lzip/clzip
1
0
Fork 0
Code Issues Activity

Merging upstream version 1.1~rc2.

Browse source 
Signed-off-by: Daniel Baumann <daniel@debian.org>
This commit is contained in:
Daniel Baumann 2025-02-17 19:14:54 +01:00
parent a8d17e4a46
commit 950a431716
Signed by: daniel
GPG key ID: FBB4F0E80A80222F
22 changed files with 1309 additions and 1071 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

7
AUTHORS
View file

@ -1,4 +1,7 @@
Clzip was written by Antonio Diaz Diaz.
Clzip implements a simplified version of the LZMA algorithm.
The original LZMA algorithm was designed by Igor Pavlov.
The ideas embodied in clzip are due to (at least) the following people:
Abraham Lempel and Jacob Ziv (for the LZ algorithm), Andrey Markov (for
the definition of Markov chains), G.N.N. Martin (for the definition of
range encoding), Igor Pavlov (for putting all the above together in
LZMA), and Julian Seward (for bzip2's CLI and the idea of unzcrash).

23
ChangeLog
View file

@ -1,3 +1,22 @@
2010-12-07 Antonio Diaz Diaz <ant_diaz@teleline.es>
* Version 1.1-rc2 released.
* main.c (open_instream): Do not show the message
" and `--stdout' was not specified" for directories, etc.
2010-08-08 Antonio Diaz Diaz <ant_diaz@teleline.es>
* Version 1.1-rc1 released.
* Code has been converted to `C89 + long long' from C99.
* main.c: Fixed warning about fchown return value being ignored.
* decoder.c: `-tvvvv' now shows compression ratio.
* main.c: Match length limit set by options -1 to -8 has been
reduced to extend range of use towards gzip. Lower numbers now
compress less but faster. (-1 now takes 43% less time for only
20% larger compressed size).
* encoder.c: Compression of option -9 has been slightly increased.
* New examples have been added to the manual.
2010-04-05 Antonio Diaz Diaz <ant_diaz@teleline.es>
* Version 1.0 released.
@ -8,5 +27,5 @@
Copyright (C) 2010 Antonio Diaz Diaz.
This file is a collection of facts, and thus it is not copyrightable,
but just in case, I give you unlimited permission to copy, distribute
and modify it.
but just in case, you have unlimited permission to copy, distribute and
modify it.

2
INSTALL
View file

@ -1,7 +1,7 @@
Requirements
------------
You will need a C compiler.
I use gcc 4.3.4 and 3.3.6, but the code should compile with any
I use gcc 4.3.5 and 3.3.6, but the code should compile with any
standards compliant compiler.
Gcc is available at http://gcc.gnu.org.

14
Makefile.in
View file

@ -16,10 +16,10 @@ objs = carg_parser.o decoder.o encoder.o main.o
all : $(progname)
$(progname) : $(objs)
$(CC) $(LDFLAGS) -o $(progname) $(objs)
$(CC) $(LDFLAGS) -o $@ $^
$(progname)_profiled : $(objs)
$(CC) $(LDFLAGS) -pg -o $(progname)_profiled $(objs)
$(CC) $(LDFLAGS) -pg -o $@ $^
main.o : main.c
$(CC) $(CPPFLAGS) $(CFLAGS) -DPROGVERSION=\"$(pkgversion)\" -c -o $@ $<
@ -44,14 +44,14 @@ $(VPATH)/doc/$(pkgname).info : $(VPATH)/doc/$(pkgname).texinfo
man : $(VPATH)/doc/$(progname).1
$(VPATH)/doc/$(progname).1 : $(progname)
help2man -n 'data compressor based on the LZMA algorithm' \
-o $(VPATH)/doc/$(progname).1 ./$(progname)
help2man -n 'reduces the size of files' \
-o $@ ./$(progname)
Makefile : $(VPATH)/configure $(VPATH)/Makefile.in
./config.status
check : all
@$(VPATH)/testsuite/check.sh $(VPATH)/testsuite
@$(VPATH)/testsuite/check.sh $(VPATH)/testsuite $(pkgversion)
install : all install-info install-man
if [ ! -d "$(DESTDIR)$(bindir)" ] ; then $(INSTALL_DIR) "$(DESTDIR)$(bindir)" ; fi
@ -94,8 +94,8 @@ dist : doc
$(DISTNAME)/doc/$(pkgname).info \
$(DISTNAME)/doc/$(pkgname).texinfo \
$(DISTNAME)/testsuite/check.sh \
$(DISTNAME)/testsuite/test1 \
$(DISTNAME)/testsuite/test1.lz \
$(DISTNAME)/testsuite/test.txt \
$(DISTNAME)/testsuite/test_v[01].lz \
$(DISTNAME)/*.h \
$(DISTNAME)/*.c
rm -f $(DISTNAME)

25
NEWS
View file

@ -1,5 +1,24 @@
Changes in version 1.0:
Changes in version 1.1:
Initial release.
Code has been converted to "C89 + long long". A C99 compiler is no more
needed.
Translated to C from the C++ source for lzip 1.10.
A warning about fchown's return value being ignored has been fixed.
"clzip -tvvvv" now shows file compression ratio.
Match length limit set by options -1 to -8 has been reduced to extend
range of use towards gzip. Lower numbers now compress less but faster.
(-1 now takes 43% less time for only 20% larger compressed size).
(Note that the bidimensional parameter space of LZMA can't be mapped to
a linear scale optimal for all files. If your files are large, very
repetitive, etc, you may need to use the --match-length and
--dictionary-size options directly to achieve optimal performance).
Compression of option -9 has been slightly increased.
Do not show the message "and `--stdout' was not specified" for file
types that can't be read (directories, etc).
Some new examples have been added to the manual.

28
README
View file

@ -35,9 +35,9 @@ standard input. This allows the direct creation of multivolume
compressed tar archives.
Clzip will automatically use the smallest possible dictionary size
without exceeding the given limit. It is important to appreciate that
the decompression memory requirement is affected at compression time by
the choice of dictionary size limit.
without exceeding the given limit. Keep in mind that the decompression
memory requirement is affected at compression time by the choice of
dictionary size limit.
As a self-check for your protection, clzip stores in the member trailer
the 32-bit CRC of the original data and the size of the original data,
@ -51,14 +51,18 @@ something is wrong. It can't help you recover the original uncompressed
data.
Clzip implements a simplified version of the LZMA (Lempel-Ziv-Markov
chain-Algorithm) algorithm. The original LZMA algorithm was designed by
Igor Pavlov.
chain-Algorithm) algorithm. The high compression of LZMA comes from
combining two basic, well-proven compression ideas: sliding dictionaries
(LZ77/78) and markov models (the thing used by every compression
algorithm that uses a range encoder or similar order-0 entropy coder as
its last stage) with segregation of contexts according to what the bits
are used for.
The high compression of LZMA comes from combining two basic, well-proven
compression ideas: sliding dictionaries (LZ77/78) and markov models (the
thing used by every compression algorithm that uses a range encoder or
similar order-0 entropy coder as its last stage) with segregation of
contexts according to what the bits are used for.
The ideas embodied in clzip are due to (at least) the following people:
Abraham Lempel and Jacob Ziv (for the LZ algorithm), Andrey Markov (for
the definition of Markov chains), G.N.N. Martin (for the definition of
range encoding), Igor Pavlov (for putting all the above together in
LZMA), and Julian Seward (for bzip2's CLI and the idea of unzcrash).
Copyright (C) 2010 Antonio Diaz Diaz.
@ -67,5 +71,5 @@ This file is free documentation: you have unlimited permission to copy,
distribute and modify it.
The file Makefile.in is a data file used by configure to produce the
Makefile. It has the same copyright owner and permissions that this
file.
Makefile. It has the same copyright owner and permissions that configure
itself.

29
carg_parser.c
View file

@ -93,16 +93,16 @@ static char parse_long_option( struct Arg_parser * const ap,
for( len = 0; opt[len+2] && opt[len+2] != '='; ++len ) ;
// Test all long options for either exact match or abbreviated matches.
/* Test all long options for either exact match or abbreviated matches. */
for( i = 0; options[i].code != 0; ++i )
if( options[i].name && !strncmp( options[i].name, &opt[2], len ) )
{
if( strlen( options[i].name ) == len ) // Exact match found
if( strlen( options[i].name ) == len ) /* Exact match found */
{ index = i; exact = 1; break; }
else if( index < 0 ) index = i; // First nonexact match found
else if( index < 0 ) index = i; /* First nonexact match found */
else if( options[index].code != options[i].code ||
options[index].has_arg != options[i].has_arg )
ambig = 1; // Second or later nonexact match found
ambig = 1; /* Second or later nonexact match found */
}
if( ambig && !exact )
@ -112,7 +112,7 @@ static char parse_long_option( struct Arg_parser * const ap,
return 1;
}
if( index < 0 ) // nothing found
if( index < 0 ) /* nothing found */
{
add_error( ap, "unrecognized option `" ); add_error( ap, opt );
add_error( ap, "'" );
@ -121,7 +121,7 @@ static char parse_long_option( struct Arg_parser * const ap,
++*argindp;
if( opt[len+2] ) // `--<long_option>=<argument>' syntax
if( opt[len+2] ) /* `--<long_option>=<argument>' syntax */
{
if( options[index].has_arg == ap_no )
{
@ -159,14 +159,15 @@ static char parse_short_option( struct Arg_parser * const ap,
const struct ap_Option options[],
int * const argindp )
{
int cind = 1; // character index in opt
int cind = 1; /* character index in opt */
while( cind > 0 )
{
int index = -1;
int i;
const unsigned char code = opt[cind];
const char code_str[2] = { code, 0 };
char code_str[2];
code_str[0] = code; code_str[1] = 0;
if( code != 0 )
for( i = 0; options[i].code; ++i )
@ -179,7 +180,7 @@ static char parse_short_option( struct Arg_parser * const ap,
return 1;
}
if( opt[++cind] == 0 ) { ++*argindp; cind = 0; } // opt finished
if( opt[++cind] == 0 ) { ++*argindp; cind = 0; } /* opt finished */
if( options[index].has_arg != ap_no && cind > 0 && opt[cind] )
{
@ -207,9 +208,9 @@ char ap_init( struct Arg_parser * const ap,
const int argc, const char * const argv[],
const struct ap_Option options[], const char in_order )
{
const char ** non_options = 0; // skipped non-options
int non_options_size = 0; // number of skipped non-options
int argind = 1; // index in argv
const char ** non_options = 0; /* skipped non-options */
int non_options_size = 0; /* number of skipped non-options */
int argind = 1; /* index in argv */
int i;
ap->data = 0;
@ -223,13 +224,13 @@ char ap_init( struct Arg_parser * const ap,
const unsigned char ch1 = argv[argind][0];
const unsigned char ch2 = ( ch1 ? argv[argind][1] : 0 );
if( ch1 == '-' && ch2 ) // we found an option
if( ch1 == '-' && ch2 ) /* we found an option */
{
const char * const opt = argv[argind];
const char * const arg = (argind + 1 < argc) ? argv[argind+1] : 0;
if( ch2 == '-' )
{
if( !argv[argind][2] ) { ++argind; break; } // we found "--"
if( !argv[argind][2] ) { ++argind; break; } /* we found "--" */
else if( !parse_long_option( ap, opt, arg, options, &argind ) ) return 0;
}
else if( !parse_short_option( ap, opt, arg, options, &argind ) ) return 0;

10
carg_parser.h
View file

@ -57,8 +57,8 @@ enum ap_Has_arg { ap_no, ap_yes, ap_maybe };
struct ap_Option
{
int code; // Short option letter or code ( code != 0 )
const char * name; // Long option name (maybe null)
int code; /* Short option letter or code ( code != 0 ) */
const char * name; /* Long option name (maybe null) */
enum ap_Has_arg has_arg;
};
@ -87,11 +87,11 @@ void ap_free( struct Arg_parser * const ap );
const char * ap_error( const struct Arg_parser * const ap );
// The number of arguments parsed (may be different from argc)
/* The number of arguments parsed (may be different from argc) */
int ap_arguments( const struct Arg_parser * const ap );
// If ap_code( i ) is 0, ap_argument( i ) is a non-option.
// Else ap_argument( i ) is the option's argument (or empty).
/* If ap_code( i ) is 0, ap_argument( i ) is a non-option.
Else ap_argument( i ) is the option's argument (or empty). */
int ap_code( const struct Arg_parser * const ap, const int i );
const char * ap_argument( const struct Arg_parser * const ap, const int i );

162
clzip.h
View file

@ -1,4 +1,4 @@
/* Clzip - A data compressor based on the LZMA algorithm
/* Clzip - Data compressor based on the LZMA algorithm
Copyright (C) 2010 Antonio Diaz Diaz.
This program is free software: you can redistribute it and/or modify
@ -24,32 +24,35 @@
typedef unsigned char State;
enum { St_states = 12 };
enum { states = 12 };
static inline bool St_is_char( const State st ) { return st < 7; }
static inline void St_set_char( State * const st )
{
static const unsigned char next[St_states] =
{0, 0, 0, 0, 1, 2, 3, 4, 5, 6, 4, 5};
static const unsigned char next[states] =
{ 0, 0, 0, 0, 1, 2, 3, 4, 5, 6, 4, 5 };
*st = next[*st];
}
static inline void St_set_match( State * const st )
{
static const unsigned char next[St_states] =
{7, 7, 7, 7, 7, 7, 7, 10, 10, 10, 10, 10};
static const unsigned char next[states] =
{ 7, 7, 7, 7, 7, 7, 7, 10, 10, 10, 10, 10 };
*st = next[*st];
}
static inline void St_set_rep( State * const st )
{
static const unsigned char next[St_states] =
{8, 8, 8, 8, 8, 8, 8, 11, 11, 11, 11, 11};
static const unsigned char next[states] =
{ 8, 8, 8, 8, 8, 8, 8, 11, 11, 11, 11, 11 };
*st = next[*st];
}
static inline void St_set_short_rep( State * const st )
{
static const unsigned char next[St_states] =
{9, 9, 9, 9, 9, 9, 9, 11, 11, 11, 11, 11};
static const unsigned char next[states] =
{ 9, 9, 9, 9, 9, 9, 9, 11, 11, 11, 11, 11 };
*st = next[*st];
}
@ -78,8 +81,8 @@ enum {
len_high_symbols = 1 << len_high_bits,
max_len_symbols = len_low_symbols + len_mid_symbols + len_high_symbols,
min_match_len = 2, // must be 2
max_match_len = min_match_len + max_len_symbols - 1, // 273
min_match_len = 2, /* must be 2 */
max_match_len = min_match_len + max_len_symbols - 1, /* 273 */
min_match_len_limit = 5,
max_dis_states = 4 };
@ -91,10 +94,10 @@ static inline int get_dis_state( int len )
return len;
}
enum {
bit_model_move_bits = 5,
bit_model_total_bits = 11,
bit_model_total = 1 << bit_model_total_bits };
enum { bit_model_move_bits = 5,
bit_model_total_bits = 11,
bit_model_total = 1 << bit_model_total_bits };
typedef unsigned int Bit_model;
@ -107,11 +110,12 @@ struct Pretty_print
const char * name_;
const char * stdin_name;
int longest_name;
int verbosity;
bool first_post;
};
void Pp_init( struct Pretty_print * const pp, const char * const filenames[],
const int num_filenames );
const int num_filenames, const int v );
static inline void Pp_set_name( struct Pretty_print * const pp,
const char * const filename )
@ -127,16 +131,18 @@ static inline void Pp_reset( struct Pretty_print * const pp )
void Pp_show_msg( struct Pretty_print * const pp, const char * const msg );
typedef uint32_t CRC32[256]; // Table of CRCs of all 8-bit messages.
typedef uint32_t CRC32[256]; /* Table of CRCs of all 8-bit messages. */
extern CRC32 crc32;
static inline void CRC32_init()
{
for( unsigned int n = 0; n < 256; ++n )
unsigned int n;
for( n = 0; n < 256; ++n )
{
unsigned int c = n;
for( int k = 0; k < 8; ++k )
int k;
for( k = 0; k < 8; ++k )
{ if( c & 1 ) c = 0xEDB88320U ^ ( c >> 1 ); else c >>= 1; }
crc32[n] = c;
}
@ -147,63 +153,66 @@ static inline void CRC32_update_byte( uint32_t * crc, const uint8_t byte )
static inline void CRC32_update_buf( uint32_t * crc, const uint8_t * const buffer,
const int size )
{
for( int i = 0; i < size; ++i )
int i;
for( i = 0; i < size; ++i )
*crc = crc32[(*crc^buffer[i])&0xFF] ^ ( *crc >> 8 );
}
typedef uint8_t File_header[6]; // 0-3 magic bytes
// 4 version
// 5 coded_dict_size;
enum { Fh_size = 6 };
static inline void Fh_set_magic( File_header header )
static inline int real_bits( const int value )
{
const uint8_t magic_string[4] = { 'L', 'Z', 'I', 'P' };
memcpy( header, magic_string, 4 );
header[4] = 1;
}
static inline bool Fh_verify_magic( const File_header header )
{
const uint8_t magic_string[4] = { 'L', 'Z', 'I', 'P' };
return ( memcmp( header, magic_string, 4 ) == 0 );
}
static inline uint8_t Fh_version( const File_header header )
{ return header[4]; }
static inline bool Fh_verify_version( const File_header header )
{ return ( header[4] <= 1 ); }
static inline int Fh_real_bits( const int value )
{
int bits = 0;
for( int i = 1, mask = 1; mask > 0; ++i, mask <<= 1 )
int bits = 0, i, mask;
for( i = 1, mask = 1; mask > 0; ++i, mask <<= 1 )
if( value & mask ) bits = i;
return bits;
}
static inline int Fh_get_dictionary_size( const File_header header )
static const uint8_t magic_string[4] = { 'L', 'Z', 'I', 'P' };
typedef uint8_t File_header[6]; /* 0-3 magic bytes */
/* 4 version */
/* 5 coded_dict_size; */
enum { Fh_size = 6 };
static inline void Fh_set_magic( File_header data )
{
int sz = ( 1 << ( header[5] & 0x1F ) );
memcpy( data, magic_string, 4 );
data[4] = 1;
}
static inline bool Fh_verify_magic( const File_header data )
{
return ( memcmp( data, magic_string, 4 ) == 0 );
}
static inline uint8_t Fh_version( const File_header data )
{ return data[4]; }
static inline bool Fh_verify_version( const File_header data )
{ return ( data[4] <= 1 ); }
static inline int Fh_get_dictionary_size( const File_header data )
{
int sz = ( 1 << ( data[5] & 0x1F ) );
if( sz > min_dictionary_size && sz <= max_dictionary_size )
sz -= ( sz / 16 ) * ( ( header[5] >> 5 ) & 0x07 );
sz -= ( sz / 16 ) * ( ( data[5] >> 5 ) & 0x07 );
return sz;
}
static inline bool Fh_set_dictionary_size( File_header header, const int sz )
static inline bool Fh_set_dictionary_size( File_header data, const int sz )
{
if( sz >= min_dictionary_size && sz <= max_dictionary_size )
{
header[5] = Fh_real_bits( sz - 1 );
data[5] = real_bits( sz - 1 );
if( sz > min_dictionary_size )
{
const int base_size = 1 << header[5];
const int base_size = 1 << data[5];
const int wedge = base_size / 16;
for( int i = 7; i >= 1; --i )
int i;
for( i = 7; i >= 1; --i )
if( base_size - ( i * wedge ) >= sz )
{ header[5] |= ( i << 5 ); break; }
{ data[5] |= ( i << 5 ); break; }
}
return true;
}
@ -212,52 +221,59 @@ static inline bool Fh_set_dictionary_size( File_header header, const int sz )
typedef uint8_t File_trailer[20];
// 0-3 CRC32 of the uncompressed data
// 4-11 size of the uncompressed data
// 12-19 member size including header and trailer
/* 0-3 CRC32 of the uncompressed data */
/* 4-11 size of the uncompressed data */
/* 12-19 member size including header and trailer */
enum { Ft_size = 20 };
static inline int Ft_versioned_size( const int version )
{ return ( ( version >= 1 ) ? 20 : 12 ); }
static inline uint32_t Ft_get_data_crc( const File_trailer trailer )
static inline uint32_t Ft_get_data_crc( const File_trailer data )
{
uint32_t tmp = 0;
for( int i = 3; i >= 0; --i ) { tmp <<= 8; tmp += trailer[i]; }
int i;
for( i = 3; i >= 0; --i ) { tmp <<= 8; tmp += data[i]; }
return tmp;
}
static inline void Ft_set_data_crc( File_trailer trailer, uint32_t crc )
{ for( int i = 0; i <= 3; ++i ) { trailer[i] = (uint8_t)crc; crc >>= 8; } }
static inline void Ft_set_data_crc( File_trailer data, uint32_t crc )
{
int i;
for( i = 0; i <= 3; ++i ) { data[i] = (uint8_t)crc; crc >>= 8; }
}
static inline long long Ft_get_data_size( const File_trailer trailer )
static inline long long Ft_get_data_size( const File_trailer data )
{
long long tmp = 0;
for( int i = 11; i >= 4; --i ) { tmp <<= 8; tmp += trailer[i]; }
int i;
for( i = 11; i >= 4; --i ) { tmp <<= 8; tmp += data[i]; }
return tmp;
}
static inline void Ft_set_data_size( File_trailer trailer, long long sz )
static inline void Ft_set_data_size( File_trailer data, long long sz )
{
for( int i = 4; i <= 11; ++i ) { trailer[i] = (uint8_t)sz; sz >>= 8; }
int i;
for( i = 4; i <= 11; ++i ) { data[i] = (uint8_t)sz; sz >>= 8; }
}
static inline long long Ft_get_member_size( const File_trailer trailer )
static inline long long Ft_get_member_size( const File_trailer data )
{
long long tmp = 0;
for( int i = 19; i >= 12; --i ) { tmp <<= 8; tmp += trailer[i]; }
int i;
for( i = 19; i >= 12; --i ) { tmp <<= 8; tmp += data[i]; }
return tmp;
}
static inline void Ft_set_member_size( File_trailer trailer, long long sz )
static inline void Ft_set_member_size( File_trailer data, long long sz )
{
for( int i = 12; i <= 19; ++i ) { trailer[i] = (uint8_t)sz; sz >>= 8; }
int i;
for( i = 12; i <= 19; ++i ) { data[i] = (uint8_t)sz; sz >>= 8; }
}
extern int verbosity;
/* defined in main.c */
void cleanup_and_fail( const int retval );
void show_error( const char * const msg, const int errcode, const bool help );
void internal_error( const char * const msg );

12
configure vendored
View file

@ -1,16 +1,16 @@
#! /bin/sh
# configure script for Clzip - A data compressor based on the LZMA algorithm
# configure script for Clzip - Data compressor based on the LZMA algorithm
# Copyright (C) 2010 Antonio Diaz Diaz.
#
# This configure script is free software: you have unlimited permission
# to copy, distribute and modify it.
#
# Date of this version: 2010-04-05
# Date of this version: 2010-12-07
args=
no_create=
pkgname=clzip
pkgversion=1.0
pkgversion=1.1-rc2
progname=clzip
srctrigger=clzip.h
@ -27,7 +27,7 @@ mandir='$(datadir)/man'
sysconfdir='$(prefix)/etc'
CC=
CPPFLAGS=
CFLAGS='-Wall -W -O2 -std=gnu99'
CFLAGS='-Wall -W -O2'
LDFLAGS=
# Loop over all args
@ -80,7 +80,7 @@ while [ -n "$1" ] ; do
bindir=`echo ${optarg} | sed -e 's,/$,,'` ;;
--datadir* | --da*)
datadir=`echo ${optarg} | sed -e 's,/$,,'` ;;
--infodir* | --in*)
--infodir* | --inf*)
infodir=`echo ${optarg} | sed -e 's,/$,,'` ;;
--mandir* | --ma*)
mandir=`echo ${optarg} | sed -e 's,/$,,'` ;;
@ -166,7 +166,7 @@ echo "CFLAGS = ${CFLAGS}"
echo "LDFLAGS = ${LDFLAGS}"
rm -f Makefile
cat > Makefile << EOF
# Makefile for Clzip - A data compressor based on the LZMA algorithm
# Makefile for Clzip - Data compressor based on the LZMA algorithm
# Copyright (C) 2010 Antonio Diaz Diaz.
# This file was generated automatically by configure. Do not edit.
#

103
decoder.c
View file

@ -1,4 +1,4 @@
/* Clzip - A data compressor based on the LZMA algorithm
/* Clzip - Data compressor based on the LZMA algorithm
Copyright (C) 2010 Antonio Diaz Diaz.
This program is free software: you can redistribute it and/or modify
@ -28,18 +28,20 @@
#include "decoder.h"
CRC32 crc32;
bool Rd_read_block( struct Range_decoder * const rdec )
{
if( !rdec->at_stream_end )
{
rdec->stream_pos = readblock( rdec->infd_, rdec->buffer, Rd_buffer_size );
if( rdec->stream_pos != Rd_buffer_size && errno )
{ show_error( "read error", errno, false ); cleanup_and_fail( 1 ); }
rdec->at_stream_end = ( rdec->stream_pos < Rd_buffer_size );
rdec->stream_pos = readblock( rdec->infd, rdec->buffer, rd_buffer_size );
if( rdec->stream_pos != rd_buffer_size && errno )
{ show_error( "Read error", errno, false ); cleanup_and_fail( 1 ); }
rdec->at_stream_end = ( rdec->stream_pos < rd_buffer_size );
rdec->partial_member_pos += rdec->pos;
rdec->pos = 0;
}
return !Rd_finished( rdec );
return rdec->pos < rdec->stream_pos;
}
@ -49,8 +51,8 @@ void LZd_flush_data( struct LZ_decoder * const decoder )
if( size > 0 )
{
CRC32_update_buf( &decoder->crc_, decoder->buffer + decoder->stream_pos, size );
if( decoder->outfd_ >= 0 &&
writeblock( decoder->outfd_, decoder->buffer + decoder->stream_pos, size ) != size )
if( decoder->outfd >= 0 &&
writeblock( decoder->outfd, decoder->buffer + decoder->stream_pos, size ) != size )
{ show_error( "write error", errno, false ); cleanup_and_fail( 1 ); }
if( decoder->pos >= decoder->buffer_size )
{ decoder->partial_data_pos += decoder->pos; decoder->pos = 0; }
@ -62,22 +64,24 @@ void LZd_flush_data( struct LZ_decoder * const decoder )
bool LZd_verify_trailer( struct LZ_decoder * const decoder,
struct Pretty_print * const pp )
{
int i;
File_trailer trailer;
const int trailer_size = Ft_versioned_size( decoder->member_version );
const long long member_size = LZd_member_position( decoder ) + trailer_size;
const long long member_size =
Rd_member_position( decoder->range_decoder ) + trailer_size;
bool error = false;
for( int i = 0; i < trailer_size && !error; ++i )
for( i = 0; i < trailer_size && !error; ++i )
{
if( !Rd_finished( decoder->range_decoder ) )
trailer[i] = Rd_get_byte( decoder->range_decoder );
else
{
error = true;
if( verbosity >= 0 )
if( pp->verbosity >= 0 )
{
Pp_show_msg( pp, 0 );
fprintf( stderr, "trailer truncated at trailer position %d;"
fprintf( stderr, "Trailer truncated at trailer position %d;"
" some checks may fail.\n", i );
}
for( ; i < trailer_size; ++i ) trailer[i] = 0;
@ -87,19 +91,15 @@ bool LZd_verify_trailer( struct LZ_decoder * const decoder,
if( !Rd_code_is_zero( decoder->range_decoder ) )
{
error = true;
if( verbosity >= 0 )
{
Pp_show_msg( pp, 0 );
fprintf( stderr, "range_decoder final code is not zero.\n" );
}
Pp_show_msg( pp, "Range decoder final code is not zero" );
}
if( Ft_get_data_crc( trailer ) != LZd_crc( decoder ) )
{
error = true;
if( verbosity >= 0 )
if( pp->verbosity >= 0 )
{
Pp_show_msg( pp, 0 );
fprintf( stderr, "crc mismatch; trailer says %08X, data crc is %08X.\n",
fprintf( stderr, "CRC mismatch; trailer says %08X, data CRC is %08X.\n",
(unsigned int)Ft_get_data_crc( trailer ),
(unsigned int)LZd_crc( decoder ) );
}
@ -107,51 +107,56 @@ bool LZd_verify_trailer( struct LZ_decoder * const decoder,
if( Ft_get_data_size( trailer ) != LZd_data_position( decoder ) )
{
error = true;
if( verbosity >= 0 )
if( pp->verbosity >= 0 )
{
Pp_show_msg( pp, 0 );
fprintf( stderr, "data size mismatch; trailer says %lld, data size is %lld (0x%llX).\n",
fprintf( stderr, "Data size mismatch; trailer says %lld, data size is %lld (0x%llX).\n",
Ft_get_data_size( trailer ), LZd_data_position( decoder ), LZd_data_position( decoder ) );
}
}
if( Ft_get_member_size( trailer ) != member_size )
{
error = true;
if( verbosity >= 0 )
if( pp->verbosity >= 0 )
{
Pp_show_msg( pp, 0 );
fprintf( stderr, "member size mismatch; trailer says %lld, member size is %lld (0x%llX).\n",
fprintf( stderr, "Member size mismatch; trailer says %lld, member size is %lld (0x%llX).\n",
Ft_get_member_size( trailer ), member_size, member_size );
}
}
if( !error && verbosity >= 3 )
fprintf( stderr, "data crc %08X, data size %9lld, member size %8lld. ",
if( !error && pp->verbosity >= 4 && LZd_data_position( decoder ) > 0 && member_size > 0 )
fprintf( stderr, "%6.3f:1, %6.3f bits/byte, %5.2f%% saved. ",
(double)LZd_data_position( decoder ) / member_size,
( 8.0 * member_size ) / LZd_data_position( decoder ),
100.0 * ( 1.0 - ( (double)member_size / LZd_data_position( decoder ) ) ) );
if( !error && pp->verbosity >= 3 )
fprintf( stderr, "data CRC %08X, data size %9lld, member size %8lld. ",
(unsigned int)Ft_get_data_crc( trailer ),
Ft_get_data_size( trailer ), Ft_get_member_size( trailer ) );
return !error;
}
// Return value: 0 = OK, 1 = decoder error, 2 = unexpected EOF,
// 3 = trailer error, 4 = unknown marker found.
/* Return value: 0 = OK, 1 = decoder error, 2 = unexpected EOF,
3 = trailer error, 4 = unknown marker found. */
int LZd_decode_member( struct LZ_decoder * const decoder,
struct Pretty_print * const pp )
{
unsigned int rep0 = 0; // rep[0-3] latest four distances
unsigned int rep1 = 0; // used for efficient coding of
unsigned int rep2 = 0; // repeated distances
unsigned int rep0 = 0; /* rep[0-3] latest four distances */
unsigned int rep1 = 0; /* used for efficient coding of */
unsigned int rep2 = 0; /* repeated distances */
unsigned int rep3 = 0;
State state = 0;
Rd_load( decoder->range_decoder );
while( true )
{
const int pos_state = LZd_data_position( decoder ) & pos_state_mask;
if( Rd_finished( decoder->range_decoder ) )
{ LZd_flush_data( decoder ); return 2; }
const int pos_state = LZd_data_position( decoder ) & pos_state_mask;
if( Rd_decode_bit( decoder->range_decoder, &decoder->bm_match[state][pos_state] ) == 0 )
{
const uint8_t prev_byte = LZd_get_byte( decoder, 0 );
const uint8_t prev_byte = LZd_get_prev_byte( decoder );
if( St_is_char( state ) )
LZd_put_byte( decoder, Lid_decode( &decoder->literal_decoder,
decoder->range_decoder, prev_byte ) );
@ -166,12 +171,7 @@ int LZd_decode_member( struct LZ_decoder * const decoder,
if( Rd_decode_bit( decoder->range_decoder, &decoder->bm_rep[state] ) == 1 )
{
len = 0;
if( Rd_decode_bit( decoder->range_decoder, &decoder->bm_rep0[state] ) == 0 )
{
if( Rd_decode_bit( decoder->range_decoder, &decoder->bm_len[state][pos_state] ) == 0 )
{ len = 1; St_set_short_rep( &state ); }
}
else
if( Rd_decode_bit( decoder->range_decoder, &decoder->bm_rep0[state] ) == 1 )
{
unsigned int distance;
if( Rd_decode_bit( decoder->range_decoder, &decoder->bm_rep1[state] ) == 0 )
@ -186,17 +186,23 @@ int LZd_decode_member( struct LZ_decoder * const decoder,
rep1 = rep0;
rep0 = distance;
}
else
{
if( Rd_decode_bit( decoder->range_decoder, &decoder->bm_len[state][pos_state] ) == 0 )
{ St_set_short_rep( &state ); len = 1; }
}
if( len == 0 )
{
len = min_match_len + Led_decode( &decoder->rep_match_len_decoder, decoder->range_decoder, pos_state );
St_set_rep( &state );
len = min_match_len + Led_decode( &decoder->rep_match_len_decoder, decoder->range_decoder, pos_state );
}
}
else
{
unsigned int rep0_saved = rep0;
int dis_slot;
const unsigned int rep0_saved = rep0;
len = min_match_len + Led_decode( &decoder->len_decoder, decoder->range_decoder, pos_state );
const int dis_slot = Rd_decode_tree( decoder->range_decoder, decoder->bm_dis_slot[get_dis_state(len)], dis_slot_bits );
dis_slot = Rd_decode_tree( decoder->range_decoder, decoder->bm_dis_slot[get_dis_state(len)], dis_slot_bits );
if( dis_slot < start_dis_model ) rep0 = dis_slot;
else
{
@ -208,32 +214,33 @@ int LZd_decode_member( struct LZ_decoder * const decoder,
{
rep0 += Rd_decode( decoder->range_decoder, direct_bits - dis_align_bits ) << dis_align_bits;
rep0 += Rd_decode_tree_reversed( decoder->range_decoder, decoder->bm_align, dis_align_bits );
if( rep0 == 0xFFFFFFFFU ) // Marker found
if( rep0 == 0xFFFFFFFFU ) /* Marker found */
{
rep0 = rep0_saved;
Rd_normalize( decoder->range_decoder );
LZd_flush_data( decoder );
if( len == min_match_len ) // End Of Stream marker
if( len == min_match_len ) /* End Of Stream marker */
{
if( LZd_verify_trailer( decoder, pp ) ) return 0; else return 3;
}
if( len == min_match_len + 1 ) // Sync Flush marker
if( len == min_match_len + 1 ) /* Sync Flush marker */
{
Rd_load( decoder->range_decoder ); continue;
}
if( verbosity >= 0 )
if( pp->verbosity >= 0 )
{
Pp_show_msg( pp, 0 );
fprintf( stderr, "unsupported marker code `%d'.\n", len );
fprintf( stderr, "Unsupported marker code `%d'.\n", len );
}
return 4;
}
if( rep0 >= (unsigned int)decoder->dictionary_size )
{ LZd_flush_data( decoder ); return 1; }
}
}
rep3 = rep2; rep2 = rep1; rep1 = rep0_saved;
St_set_match( &state );
if( rep0 >= (unsigned int)decoder->dictionary_size ||
( rep0 >= (unsigned int)decoder->pos && !decoder->partial_data_pos ) )
{ LZd_flush_data( decoder ); return 1; }
}
LZd_copy_block( decoder, rep0, len );
}

134
decoder.h
View file

@ -1,4 +1,4 @@
/* Clzip - A data compressor based on the LZMA algorithm
/* Clzip - Data compressor based on the LZMA algorithm
Copyright (C) 2010 Antonio Diaz Diaz.
This program is free software: you can redistribute it and/or modify
@ -15,36 +15,36 @@
along with this program. If not, see <http://www.gnu.org/licenses/>.
*/
enum { Rd_buffer_size = 16384 };
enum { rd_buffer_size = 16384 };
struct Range_decoder
{
long long partial_member_pos;
uint8_t * buffer; // input buffer
int pos;
int stream_pos; // when reached, a new block must be read
uint8_t * buffer; /* input buffer */
int pos; /* current pos in buffer */
int stream_pos; /* when reached, a new block must be read */
uint32_t code;
uint32_t range;
int infd_; // input file descriptor
int infd; /* input file descriptor */
bool at_stream_end;
};
bool Rd_read_block( struct Range_decoder * const rdec );
static inline void Rd_init( struct Range_decoder * const rdec, const int infd )
static inline void Rd_init( struct Range_decoder * const rdec, const int ifd )
{
rdec->partial_member_pos = 0;
rdec->buffer = (uint8_t *)malloc( Rd_buffer_size );
rdec->buffer = (uint8_t *)malloc( rd_buffer_size );
if( !rdec->buffer )
{
show_error( "not enough memory. Find a machine with more memory", 0, false );
show_error( "Not enough memory. Find a machine with more memory.", 0, false );
cleanup_and_fail( 1 );
}
rdec->pos = 0;
rdec->stream_pos = 0;
rdec->code = 0;
rdec->range = 0xFFFFFFFFU;
rdec->infd_ = infd;
rdec->infd = ifd;
rdec->at_stream_end = false;
}
@ -55,7 +55,7 @@ static inline bool Rd_code_is_zero( struct Range_decoder * const rdec )
{ return ( rdec->code == 0 ); }
static inline bool Rd_finished( struct Range_decoder * const rdec )
{ return rdec->at_stream_end && rdec->pos >= rdec->stream_pos; }
{ return rdec->pos >= rdec->stream_pos && !Rd_read_block( rdec ); }
static inline long long Rd_member_position( struct Range_decoder * const rdec )
{ return rdec->partial_member_pos + rdec->pos; }
@ -65,15 +65,16 @@ static inline void Rd_reset_member_position( struct Range_decoder * const rdec )
static inline uint8_t Rd_get_byte( struct Range_decoder * const rdec )
{
if( rdec->pos >= rdec->stream_pos && !Rd_read_block( rdec ) ) return 0;
if( Rd_finished( rdec ) ) return 0;
return rdec->buffer[rdec->pos++];
}
static inline void Rd_load( struct Range_decoder * const rdec )
{
int i;
rdec->code = 0;
rdec->range = 0xFFFFFFFFU;
for( int i = 0; i < 5; ++i )
for( i = 0; i < 5; ++i )
rdec->code = (rdec->code << 8) | Rd_get_byte( rdec );
}
@ -90,7 +91,8 @@ static inline int Rd_decode( struct Range_decoder * const rdec,
const int num_bits )
{
int symbol = 0;
for( int i = num_bits; i > 0; --i )
int i;
for( i = num_bits; i > 0; --i )
{
symbol <<= 1;
if( rdec->range <= 0x00FFFFFFU )
@ -113,9 +115,9 @@ static inline int Rd_decode( struct Range_decoder * const rdec,
static inline int Rd_decode_bit( struct Range_decoder * const rdec,
Bit_model * const probability )
{
uint32_t bound;
Rd_normalize( rdec );
const uint32_t bound = ( rdec->range >> bit_model_total_bits ) *
*probability;
bound = ( rdec->range >> bit_model_total_bits ) * *probability;
if( rdec->code < bound )
{
rdec->range = bound;
@ -135,7 +137,8 @@ static inline int Rd_decode_tree( struct Range_decoder * const rdec,
Bit_model bm[], const int num_bits )
{
int model = 1;
for( int i = num_bits; i > 0; --i )
int i;
for( i = num_bits; i > 0; --i )
model = ( model << 1 ) | Rd_decode_bit( rdec, &bm[model] );
return model - (1 << num_bits);
}
@ -145,7 +148,8 @@ static inline int Rd_decode_tree_reversed( struct Range_decoder * const rdec,
{
int model = 1;
int symbol = 0;
for( int i = 0; i < num_bits; ++i )
int i;
for( i = 0; i < num_bits; ++i )
{
const int bit = Rd_decode_bit( rdec, &bm[model] );
model <<= 1;
@ -159,7 +163,8 @@ static inline int Rd_decode_matched( struct Range_decoder * const rdec,
{
Bit_model * const bm1 = bm + 0x100;
int symbol = 1;
for( int i = 7; i >= 0; --i )
int i;
for( i = 7; i >= 0; --i )
{
const int match_bit = ( match_byte >> i ) & 1;
const int bit = Rd_decode_bit( rdec, &bm1[(match_bit<<8)+symbol] );
@ -186,15 +191,16 @@ struct Len_decoder
static inline void Led_init( struct Len_decoder * const len_decoder )
{
int i, j;
Bm_init( &len_decoder->choice1 );
Bm_init( &len_decoder->choice2 );
for( int i = 0; i < pos_states; ++i )
for( int j = 0; j < len_low_symbols; ++j )
for( i = 0; i < pos_states; ++i )
for( j = 0; j < len_low_symbols; ++j )
Bm_init( &len_decoder->bm_low[i][j] );
for( int i = 0; i < pos_states; ++i )
for( int j = 0; j < len_mid_symbols; ++j )
for( i = 0; i < pos_states; ++i )
for( j = 0; j < len_mid_symbols; ++j )
Bm_init( &len_decoder->bm_mid[i][j] );
for( int i = 0; i < len_high_symbols; ++i )
for( i = 0; i < len_high_symbols; ++i )
Bm_init( &len_decoder->bm_high[i] );
}
@ -219,8 +225,9 @@ struct Literal_decoder
static inline void Lid_init( struct Literal_decoder * const literal_decoder )
{
for( int i = 0; i < 1<<literal_context_bits; ++i )
for( int j = 0; j < 0x300; ++j )
int i, j;
for( i = 0; i < 1<<literal_context_bits; ++i )
for( j = 0; j < 0x300; ++j )
Bm_init( &literal_decoder->bm_literal[i][j] );
}
@ -242,23 +249,23 @@ static inline uint8_t Lid_decode_matched( struct Literal_decoder * const literal
struct LZ_decoder
{
long long partial_data_pos;
int member_version;
int dictionary_size;
int buffer_size;
uint8_t * buffer;
int pos;
int stream_pos; // first byte not yet written to file
uint8_t * buffer; /* output buffer */
int pos; /* current pos in buffer */
int stream_pos; /* first byte not yet written to file */
uint32_t crc_;
int outfd_; // output file descriptor
int outfd; /* output file descriptor */
int member_version;
Bit_model bm_match[St_states][pos_states];
Bit_model bm_rep[St_states];
Bit_model bm_rep0[St_states];
Bit_model bm_rep1[St_states];
Bit_model bm_rep2[St_states];
Bit_model bm_len[St_states][pos_states];
Bit_model bm_match[states][pos_states];
Bit_model bm_rep[states];
Bit_model bm_rep0[states];
Bit_model bm_rep1[states];
Bit_model bm_rep2[states];
Bit_model bm_len[states][pos_states];
Bit_model bm_dis_slot[max_dis_states][1<<dis_slot_bits];
Bit_model bm_dis[modeled_distances-end_dis_model];
Bit_model bm_dis[modeled_distances-end_dis_model+1];
Bit_model bm_align[dis_align_size];
struct Range_decoder * range_decoder;
@ -269,6 +276,16 @@ struct LZ_decoder
void LZd_flush_data( struct LZ_decoder * const decoder );
bool LZd_verify_trailer( struct LZ_decoder * const decoder,
struct Pretty_print * const pp );
static inline uint8_t LZd_get_prev_byte( struct LZ_decoder * const decoder )
{
const int i =
( ( decoder->pos > 0 ) ? decoder->pos : decoder->buffer_size ) - 1;
return decoder->buffer[i];
}
static inline uint8_t LZd_get_byte( struct LZ_decoder * const decoder,
const int distance )
{
@ -281,8 +298,7 @@ static inline void LZd_put_byte( struct LZ_decoder * const decoder,
const uint8_t b )
{
decoder->buffer[decoder->pos] = b;
if( ++decoder->pos >= decoder->buffer_size )
LZd_flush_data( decoder );
if( ++decoder->pos >= decoder->buffer_size ) LZd_flush_data( decoder );
}
static inline void LZd_copy_block( struct LZ_decoder * const decoder,
@ -299,37 +315,34 @@ static inline void LZd_copy_block( struct LZ_decoder * const decoder,
else for( ; len > 0 ; --len )
{
decoder->buffer[decoder->pos] = decoder->buffer[i];
if( ++decoder->pos >= decoder->buffer_size )
LZd_flush_data( decoder );
if( ++decoder->pos >= decoder->buffer_size ) LZd_flush_data( decoder );
if( ++i >= decoder->buffer_size ) i = 0;
}
}
bool LZd_verify_trailer( struct LZ_decoder * const decoder,
struct Pretty_print * const pp );
static inline void LZd_init( struct LZ_decoder * const decoder,
const File_header header,
struct Range_decoder * const rdec, const int outfd )
struct Range_decoder * const rdec, const int ofd )
{
int i, j;
decoder->partial_data_pos = 0;
decoder->member_version = Fh_version( header );
decoder->dictionary_size = Fh_get_dictionary_size( header );
decoder->buffer_size = max( 65536, decoder->dictionary_size );
decoder->buffer = (uint8_t *)malloc( decoder->buffer_size );
if( !decoder->buffer )
{
show_error( "not enough memory. Find a machine with more memory", 0, false );
show_error( "Not enough memory. Find a machine with more memory.", 0, false );
cleanup_and_fail( 1 );
}
decoder->pos = 0;
decoder->stream_pos = 0;
decoder->crc_ = 0xFFFFFFFFU;
decoder->outfd_ = outfd;
decoder->outfd = ofd;
decoder->member_version = Fh_version( header );
for( int i = 0; i < St_states; ++i )
for( i = 0; i < states; ++i )
{
for( int j = 0; j < pos_states; ++j )
for( j = 0; j < pos_states; ++j )
{
Bm_init( &decoder->bm_match[i][j] );
Bm_init( &decoder->bm_len[i][j] );
@ -339,19 +352,19 @@ static inline void LZd_init( struct LZ_decoder * const decoder,
Bm_init( &decoder->bm_rep1[i] );
Bm_init( &decoder->bm_rep2[i] );
}
for( int i = 0; i < max_dis_states; ++i )
for( int j = 0; j < 1<<dis_slot_bits; ++j )
for( i = 0; i < max_dis_states; ++i )
for( j = 0; j < 1<<dis_slot_bits; ++j )
Bm_init( &decoder->bm_dis_slot[i][j] );
for( int i = 0; i < modeled_distances-end_dis_model; ++i )
for( i = 0; i < modeled_distances-end_dis_model+1; ++i )
Bm_init( &decoder->bm_dis[i] );
for( int i = 0; i < dis_align_size; ++i )
for( i = 0; i < dis_align_size; ++i )
Bm_init( &decoder->bm_align[i] );
decoder->range_decoder = rdec;
Led_init( &decoder->len_decoder );
Led_init( &decoder->rep_match_len_decoder );
Lid_init( &decoder->literal_decoder );
decoder->buffer[decoder->buffer_size-1] = 0; // prev_byte of first_byte
decoder->buffer[decoder->buffer_size-1] = 0; /* prev_byte of first_byte */
}
static inline void LZd_free( struct LZ_decoder * const decoder )
@ -360,11 +373,8 @@ static inline void LZd_free( struct LZ_decoder * const decoder )
static inline uint32_t LZd_crc( struct LZ_decoder * const decoder )
{ return decoder->crc_ ^ 0xFFFFFFFFU; }
int LZd_decode_member( struct LZ_decoder * const decoder,
struct Pretty_print * const pp );
static inline long long LZd_member_position( struct LZ_decoder * const decoder )
{ return Rd_member_position( decoder->range_decoder ); }
static inline long long LZd_data_position( struct LZ_decoder * const decoder )
{ return decoder->partial_data_pos + decoder->pos; }
int LZd_decode_member( struct LZ_decoder * const decoder,
struct Pretty_print * const pp );

6
doc/clzip.1
View file

@ -1,12 +1,12 @@
.\" DO NOT MODIFY THIS FILE! It was generated by help2man 1.37.1.
.TH CLZIP "1" "April 2010" "Clzip 1.0" "User Commands"
.TH CLZIP "1" "December 2010" "Clzip 1.1-rc2" "User Commands"
.SH NAME
Clzip \- data compressor based on the LZMA algorithm
Clzip \- reduces the size of files
.SH SYNOPSIS
.B clzip
[\fIoptions\fR] [\fIfiles\fR]
.SH DESCRIPTION
Clzip \- A data compressor based on the LZMA algorithm.
Clzip \- Data compressor based on the LZMA algorithm.
.SH OPTIONS
.TP
\fB\-h\fR, \fB\-\-help\fR

168
doc/clzip.info
View file

@ -12,17 +12,17 @@ File: clzip.info, Node: Top, Next: Introduction, Up: (dir)
Clzip Manual
************
This manual is for Clzip (version 1.0, 5 April 2010).
This manual is for Clzip (version 1.1-rc2, 7 December 2010).
* Menu:
* Introduction:: Purpose and features of clzip
* Algorithm:: How clzip compresses the data
* Invoking Clzip:: Command line interface
* File Format:: Detailed format of the compressed file
* Examples:: A small tutorial with examples
* Problems:: Reporting bugs
* Concept Index:: Index of concepts
* Introduction:: Purpose and features of clzip
* Algorithm:: How clzip compresses the data
* Invoking Clzip:: Command line interface
* File Format:: Detailed format of the compressed file
* Examples:: A small tutorial with examples
* Problems:: Reporting bugs
* Concept Index:: Index of concepts
Copyright (C) 2010 Antonio Diaz Diaz.
@ -74,10 +74,12 @@ multivolume compressed tar archives.
or 2 times the dictionary size limit (1 if input file size is less than
dictionary size limit, else 2) plus 8 times the dictionary size really
used. For decompression it is a little more than the dictionary size
really used. Clzip will automatically use the smallest possible
dictionary size without exceeding the given limit. It is important to
appreciate that the decompression memory requirement is affected at
compression time by the choice of dictionary size limit.
really used.
Clzip will automatically use the smallest possible dictionary size
without exceeding the given limit. Keep in mind that the decompression
memory requirement is affected at compression time by the choice of
dictionary size limit.
When decompressing, clzip attempts to guess the name for the
decompressed file from that of the compressed file as follows:
@ -109,14 +111,12 @@ File: clzip.info, Node: Algorithm, Next: Invoking Clzip, Prev: Introduction,
***********
Clzip implements a simplified version of the LZMA (Lempel-Ziv-Markov
chain-Algorithm) algorithm. The original LZMA algorithm was designed by
Igor Pavlov.
The high compression of LZMA comes from combining two basic,
well-proven compression ideas: sliding dictionaries (LZ77/78) and
markov models (the thing used by every compression algorithm that uses
a range encoder or similar order-0 entropy coder as its last stage)
with segregation of contexts according to what the bits are used for.
chain-Algorithm) algorithm. The high compression of LZMA comes from
combining two basic, well-proven compression ideas: sliding dictionaries
(LZ77/78) and markov models (the thing used by every compression
algorithm that uses a range encoder or similar order-0 entropy coder as
its last stage) with segregation of contexts according to what the bits
are used for.
Clzip is a two stage compressor. The first stage is a Lempel-Ziv
coder, which reduces redundancy by translating chunks of data to their
@ -158,6 +158,13 @@ member or volume size limits are reached.
10) If there are more data to compress, go back to step 1.
The ideas embodied in clzip are due to (at least) the following people:
Abraham Lempel and Jacob Ziv (for the LZ algorithm), Andrey Markov (for
the definition of Markov chains), G.N.N. Martin (for the definition of
range encoding), Igor Pavlov (for putting all the above together in
LZMA), and Julian Seward (for bzip2's CLI and the idea of unzcrash).

File: clzip.info, Node: Invoking Clzip, Next: File Format, Prev: Algorithm, Up: Top
@ -170,49 +177,50 @@ The format for running clzip is:
Clzip supports the following options:
`--help'
`-h'
`--help'
Print an informative help message describing the options and exit.
`--version'
`-V'
`--version'
Print the version number of clzip on the standard output and exit.
`--member-size=SIZE'
`-b SIZE'
`--member-size=SIZE'
Produce a multimember file and set the member size limit to SIZE
bytes. Minimum member size limit is 100kB. Small member size may
degrade compression ratio, so use it only when needed. The default
is to produce single member files.
is to produce single-member files.
`--stdout'
`-c'
`--stdout'
Compress or decompress to standard output. Needed when reading
from a named pipe (fifo) or from a device. Use it to recover as
much of the uncompressed data as possible when decompressing a
corrupt file.
`--decompress'
`-d'
`--decompress'
Decompress.
`--force'
`-f'
`--force'
Force overwrite of output file.
`--keep'
`-k'
`--keep'
Keep (don't delete) input files during compression or
decompression.
`--match-length=LENGTH'
`-m LENGTH'
Set the match length limit in bytes. Valid values range from 5 to
273. Larger values usually give better compression ratios but
longer compression times.
`--match-length=LENGTH'
Set the match length limit in bytes. After a match this long is
found, the search is finished. Valid values range from 5 to 273.
Larger values usually give better compression ratios but longer
compression times.
`--output=FILE'
`-o FILE'
`--output=FILE'
When reading from standard input and `--stdout' has not been
specified, use `FILE' as the virtual name of the uncompressed
file. This produces a file named `FILE' when decompressing, a file
@ -220,20 +228,25 @@ The format for running clzip is:
`FILE00001.lz', `FILE00002.lz', etc, when compressing and
splitting the output in volumes.
`--quiet'
`-q'
`--quiet'
Quiet operation. Suppress all messages.
`--dictionary-size=SIZE'
`-s SIZE'
`--dictionary-size=SIZE'
Set the dictionary size limit in bytes. Valid values range from
4KiB to 512MiB. Clzip will use the smallest possible dictionary
size for each member without exceeding this limit. Note that
dictionary sizes are quantized. If the specified size does not
match one of the valid sizes, it will be rounded upwards.
`--volume-size=SIZE'
For maximum compression you should use a dictionary size limit as
large as possible, but keep in mind that the decompression memory
requirement is affected at compression time by the choice of
dictionary size limit.
`-S SIZE'
`--volume-size=SIZE'
Split the compressed output into several volume files with names
`original_name00001.lz', `original_name00002.lz', etc, and set the
volume size limit to SIZE bytes. Each volume is a complete, maybe
@ -241,15 +254,15 @@ The format for running clzip is:
volume size may degrade compression ratio, so use it only when
needed.
`--test'
`-t'
`--test'
Check integrity of the specified file(s), but don't decompress
them. This really performs a trial decompression and throws away
the result. Use `-tvv' or `-tvvv' to see information about the
file.
the result. Use it together with `-v' to see information about
the file.
`--verbose'
`-v'
`--verbose'
Verbose mode. Show the compression ratio for each file processed.
Further -v's increase the verbosity level.
@ -258,15 +271,21 @@ The format for running clzip is:
limit) as shown in the table below. Note that `-9' can be much
slower than `-1'. These options have no effect when decompressing.
The bidimensional parameter space of LZMA can't be mapped to a
linear scale optimal for all files. If your files are large, very
repetitive, etc, you may need to use the `--match-length' and
`--dictionary-size' options directly to achieve optimal
performance.
Level Dictionary size Match length limit
-1 1 MiB 10 bytes
-2 1.5 MiB 12 bytes
-3 2 MiB 17 bytes
-4 3 MiB 26 bytes
-5 4 MiB 44 bytes
-6 8 MiB 80 bytes
-7 16 MiB 108 bytes
-8 24 MiB 163 bytes
-1 1 MiB 5 bytes
-2 1.5 MiB 6 bytes
-3 2 MiB 8 bytes
-4 3 MiB 12 bytes
-5 4 MiB 20 bytes
-6 8 MiB 36 bytes
-7 16 MiB 68 bytes
-8 24 MiB 132 bytes
-9 32 MiB 273 bytes
`--fast'
@ -321,7 +340,7 @@ additional information before, between, or after them.
All multibyte values are stored in little endian order.
`ID string'
A four byte string, identifying the member type, with the value
A four byte string, identifying the lzip format, with the value
"LZIP".
`VN (version number, 1 byte)'
@ -358,9 +377,12 @@ File: clzip.info, Node: Examples, Next: Problems, Prev: File Format, Up: Top
5 A small tutorial with examples
********************************
WARNING! If your data is important, give the `--keep' option to clzip
and do not remove the original file until you verify the compressed
file with a command like `clzip -cd file.lz | cmp file -'.
WARNING! Even if clzip is bug-free, other causes may result in a corrupt
compressed file (bugs in the system libraries, memory errors, etc).
Therefore, if the data you are going to compress is important give the
`--keep' option to clzip and do not remove the original file until you
verify the compressed file with a command like
`clzip -cd file.lz | cmp file -'.
Example 1: Replace a regular file with its compressed version file.lz
@ -370,29 +392,47 @@ and show the compression ratio.
Example 2: Like example 1 but the created file.lz is multimember with a
member size of 1MiB.
member size of 1MiB. The compression ratio is not shown.
clzip -b 1MiB file
Example 3: Compress a whole floppy in /dev/fd0 and send the output to
Example 3: Restore a regular file from its compressed version file.lz.
If the operation is successful, file.lz is removed.
clzip -d file.lz
Example 4: Verify the integrity of the compressed file file.lz and show
status.
clzip -tv file.lz
Example 5: Compress a whole floppy in /dev/fd0 and send the output to
file.lz.
clzip -c /dev/fd0 > file.lz
Example 4: Create a multivolume compressed tar archive with a volume
Example 6: Decompress file.lz partially until 10KiB of decompressed data
are produced.
clzip -cd file.lz | dd bs=1024 count=10
Example 7: Create a multivolume compressed tar archive with a volume
size of 1440KiB.
tar -c some_directory | clzip -S 1440KiB -o volume_name
Example 5: Extract a multivolume compressed tar archive.
Example 8: Extract a multivolume compressed tar archive.
clzip -cd volume_name*.lz | tar -xf -
Example 6: Create a multivolume compressed backup of a big database file
Example 9: Create a multivolume compressed backup of a big database file
with a volume size of 650MB, where each volume is a multimember file
with a member size of 32MiB.
@ -437,12 +477,12 @@ Concept Index

Tag Table:
Node: Top226
Node: Introduction830
Node: Algorithm4377
Node: Invoking Clzip6608
Node: File Format10964
Node: Examples12920
Node: Problems14097
Node: Concept Index14623
Node: Introduction905
Node: Algorithm4439
Node: Invoking Clzip6963
Node: File Format11904
Node: Examples13860
Node: Problems15629
Node: Concept Index16155

End Tag Table

194
doc/clzip.texinfo
View file

@ -5,8 +5,8 @@
@finalout
@c %**end of header
@set UPDATED 5 April 2010
@set VERSION 1.0
@set UPDATED 7 December 2010
@set VERSION 1.1-rc2
@dircategory Data Compression
@direntry
@ -14,9 +14,10 @@
@end direntry
@ifnothtml
@titlepage
@title Clzip
@subtitle A data compressor based on the LZMA algorithm
@subtitle Data compressor based on the LZMA algorithm
@subtitle for Clzip version @value{VERSION}, @value{UPDATED}
@author by Antonio Diaz Diaz
@ -25,6 +26,7 @@
@end titlepage
@contents
@end ifnothtml
@node Top
@top
@ -32,13 +34,13 @@
This manual is for Clzip (version @value{VERSION}, @value{UPDATED}).
@menu
* Introduction:: Purpose and features of clzip
* Algorithm:: How clzip compresses the data
* Invoking Clzip:: Command line interface
* File Format:: Detailed format of the compressed file
* Examples:: A small tutorial with examples
* Problems:: Reporting bugs
* Concept Index:: Index of concepts
* Introduction:: Purpose and features of clzip
* Algorithm:: How clzip compresses the data
* Invoking Clzip:: Command line interface
* File Format:: Detailed format of the compressed file
* Examples:: A small tutorial with examples
* Problems:: Reporting bugs
* Concept Index:: Index of concepts
@end menu
@sp 1
@ -90,10 +92,12 @@ The amount of memory required for compression is about 5 MiB plus 1 or 2
times the dictionary size limit (1 if input file size is less than
dictionary size limit, else 2) plus 8 times the dictionary size really
used. For decompression it is a little more than the dictionary size
really used. Clzip will automatically use the smallest possible
dictionary size without exceeding the given limit. It is important to
appreciate that the decompression memory requirement is affected at
compression time by the choice of dictionary size limit.
really used.
Clzip will automatically use the smallest possible dictionary size
without exceeding the given limit. Keep in mind that the decompression
memory requirement is affected at compression time by the choice of
dictionary size limit.
When decompressing, clzip attempts to guess the name for the decompressed
file from that of the compressed file as follows:
@ -126,14 +130,12 @@ caused clzip to panic.
@cindex algorithm
Clzip implements a simplified version of the LZMA (Lempel-Ziv-Markov
chain-Algorithm) algorithm. The original LZMA algorithm was designed by
Igor Pavlov.
The high compression of LZMA comes from combining two basic, well-proven
compression ideas: sliding dictionaries (LZ77/78) and markov models (the
thing used by every compression algorithm that uses a range encoder or
similar order-0 entropy coder as its last stage) with segregation of
contexts according to what the bits are used for.
chain-Algorithm) algorithm. The high compression of LZMA comes from
combining two basic, well-proven compression ideas: sliding dictionaries
(LZ77/78) and markov models (the thing used by every compression
algorithm that uses a range encoder or similar order-0 entropy coder as
its last stage) with segregation of contexts according to what the bits
are used for.
Clzip is a two stage compressor. The first stage is a Lempel-Ziv coder,
which reduces redundancy by translating chunks of data to their
@ -175,6 +177,14 @@ member or volume size limits are reached.
10) If there are more data to compress, go back to step 1.
@sp 1
@noindent
The ideas embodied in clzip are due to (at least) the following people:
Abraham Lempel and Jacob Ziv (for the LZ algorithm), Andrey Markov (for
the definition of Markov chains), G.N.N. Martin (for the definition of
range encoding), Igor Pavlov (for putting all the above together in
LZMA), and Julian Seward (for bzip2's CLI and the idea of unzcrash).
@node Invoking Clzip
@chapter Invoking Clzip
@ -192,47 +202,47 @@ clzip [@var{options}] [@var{files}]
Clzip supports the following options:
@table @samp
@item --help
@itemx -h
@item -h
@itemx --help
Print an informative help message describing the options and exit.
@item --version
@itemx -V
@item -V
@itemx --version
Print the version number of clzip on the standard output and exit.
@item --member-size=@var{size}
@itemx -b @var{size}
@item -b @var{size}
@itemx --member-size=@var{size}
Produce a multimember file and set the member size limit to @var{size}
bytes. Minimum member size limit is 100kB. Small member size may degrade
compression ratio, so use it only when needed. The default is to produce
single member files.
single-member files.
@item --stdout
@itemx -c
@item -c
@itemx --stdout
Compress or decompress to standard output. Needed when reading from a
named pipe (fifo) or from a device. Use it to recover as much of the
uncompressed data as possible when decompressing a corrupt file.
@item --decompress
@itemx -d
@item -d
@itemx --decompress
Decompress.
@item --force
@itemx -f
@item -f
@itemx --force
Force overwrite of output file.
@item --keep
@itemx -k
@item -k
@itemx --keep
Keep (don't delete) input files during compression or decompression.
@item --match-length=@var{length}
@itemx -m @var{length}
Set the match length limit in bytes. Valid values range from 5 to 273.
Larger values usually give better compression ratios but longer
compression times.
@item -m @var{length}
@itemx --match-length=@var{length}
Set the match length limit in bytes. After a match this long is found,
the search is finished. Valid values range from 5 to 273. Larger values
usually give better compression ratios but longer compression times.
@item --output=@var{file}
@itemx -o @var{file}
@item -o @var{file}
@itemx --output=@var{file}
When reading from standard input and @samp{--stdout} has not been
specified, use @samp{@var{file}} as the virtual name of the uncompressed
file. This produces a file named @samp{@var{file}} when decompressing, a
@ -240,34 +250,38 @@ file named @samp{@var{file}.lz} when compressing, and several files
named @samp{@var{file}00001.lz}, @samp{@var{file}00002.lz}, etc, when
compressing and splitting the output in volumes.
@item --quiet
@itemx -q
@item -q
@itemx --quiet
Quiet operation. Suppress all messages.
@item --dictionary-size=@var{size}
@itemx -s @var{size}
@item -s @var{size}
@itemx --dictionary-size=@var{size}
Set the dictionary size limit in bytes. Valid values range from 4KiB to
512MiB. Clzip will use the smallest possible dictionary size for each
member without exceeding this limit. Note that dictionary sizes are
quantized. If the specified size does not match one of the valid sizes,
it will be rounded upwards.
@item --volume-size=@var{size}
@itemx -S @var{size}
For maximum compression you should use a dictionary size limit as large
as possible, but keep in mind that the decompression memory requirement
is affected at compression time by the choice of dictionary size limit.
@item -S @var{size}
@itemx --volume-size=@var{size}
Split the compressed output into several volume files with names
@samp{original_name00001.lz}, @samp{original_name00002.lz}, etc, and set
the volume size limit to @var{size} bytes. Each volume is a complete,
maybe multimember, lzip file. Minimum volume size limit is 100kB. Small
volume size may degrade compression ratio, so use it only when needed.
@item --test
@itemx -t
@item -t
@itemx --test
Check integrity of the specified file(s), but don't decompress them.
This really performs a trial decompression and throws away the result.
Use @samp{-tvv} or @samp{-tvvv} to see information about the file.
Use it together with @samp{-v} to see information about the file.
@item --verbose
@itemx -v
@item -v
@itemx --verbose
Verbose mode. Show the compression ratio for each file processed.
Further -v's increase the verbosity level.
@ -276,16 +290,22 @@ Set the compression parameters (dictionary size and match length limit)
as shown in the table below. Note that @samp{-9} can be much slower than
@samp{-1}. These options have no effect when decompressing.
The bidimensional parameter space of LZMA can't be mapped to a linear
scale optimal for all files. If your files are large, very repetitive,
etc, you may need to use the @samp{--match-length} and
@samp{--dictionary-size} options directly to achieve optimal
performance.
@multitable {Level} {Dictionary size} {Match length limit}
@item Level @tab Dictionary size @tab Match length limit
@item -1 @tab 1 MiB @tab 10 bytes
@item -2 @tab 1.5 MiB @tab 12 bytes
@item -3 @tab 2 MiB @tab 17 bytes
@item -4 @tab 3 MiB @tab 26 bytes
@item -5 @tab 4 MiB @tab 44 bytes
@item -6 @tab 8 MiB @tab 80 bytes
@item -7 @tab 16 MiB @tab 108 bytes
@item -8 @tab 24 MiB @tab 163 bytes
@item -1 @tab 1 MiB @tab 5 bytes
@item -2 @tab 1.5 MiB @tab 6 bytes
@item -3 @tab 2 MiB @tab 8 bytes
@item -4 @tab 3 MiB @tab 12 bytes
@item -5 @tab 4 MiB @tab 20 bytes
@item -6 @tab 8 MiB @tab 36 bytes
@item -7 @tab 16 MiB @tab 68 bytes
@item -8 @tab 24 MiB @tab 132 bytes
@item -9 @tab 32 MiB @tab 273 bytes
@end multitable
@ -350,7 +370,7 @@ All multibyte values are stored in little endian order.
@table @samp
@item ID string
A four byte string, identifying the member type, with the value "LZIP".
A four byte string, identifying the lzip format, with the value "LZIP".
@item VN (version number, 1 byte)
Just in case something needs to be modified in the future. Valid values
@ -385,9 +405,12 @@ safe recovery of undamaged members from multimember files.
@chapter A small tutorial with examples
@cindex examples
WARNING! If your data is important, give the @samp{--keep} option to
clzip and do not remove the original file until you verify the compressed
file with a command like @samp{clzip -cd file.lz | cmp file -}.
WARNING! Even if clzip is bug-free, other causes may result in a corrupt
compressed file (bugs in the system libraries, memory errors, etc).
Therefore, if the data you are going to compress is important give the
@samp{--keep} option to clzip and do not remove the original file until
you verify the compressed file with a command like @w{@samp{clzip -cd
file.lz | cmp file -}}.
@sp 1
@noindent
@ -401,7 +424,7 @@ clzip -v file
@sp 1
@noindent
Example 2: Like example 1 but the created file.lz is multimember with a
member size of 1MiB.
member size of 1MiB. The compression ratio is not shown.
@example
clzip -b 1MiB file
@ -409,7 +432,25 @@ clzip -b 1MiB file
@sp 1
@noindent
Example 3: Compress a whole floppy in /dev/fd0 and send the output to
Example 3: Restore a regular file from its compressed version file.lz.
If the operation is successful, file.lz is removed.
@example
clzip -d file.lz
@end example
@sp 1
@noindent
Example 4: Verify the integrity of the compressed file file.lz and show
status.
@example
clzip -tv file.lz
@end example
@sp 1
@noindent
Example 5: Compress a whole floppy in /dev/fd0 and send the output to
file.lz.
@example
@ -418,7 +459,16 @@ clzip -c /dev/fd0 > file.lz
@sp 1
@noindent
Example 4: Create a multivolume compressed tar archive with a volume
Example 6: Decompress file.lz partially until 10KiB of decompressed data
are produced.
@example
clzip -cd file.lz | dd bs=1024 count=10
@end example
@sp 1
@noindent
Example 7: Create a multivolume compressed tar archive with a volume
size of 1440KiB.
@example
@ -427,7 +477,7 @@ tar -c some_directory | clzip -S 1440KiB -o volume_name
@sp 1
@noindent
Example 5: Extract a multivolume compressed tar archive.
Example 8: Extract a multivolume compressed tar archive.
@example
clzip -cd volume_name*.lz | tar -xf -
@ -435,7 +485,7 @@ clzip -cd volume_name*.lz | tar -xf -
@sp 1
@noindent
Example 6: Create a multivolume compressed backup of a big database file
Example 9: Create a multivolume compressed backup of a big database file
with a volume size of 650MB, where each volume is a multimember file
with a member size of 32MiB.

433
encoder.c
View file

@ -1,4 +1,4 @@
/* Clzip - A data compressor based on the LZMA algorithm
/* Clzip - Data compressor based on the LZMA algorithm
Copyright (C) 2010 Antonio Diaz Diaz.
This program is free software: you can redistribute it and/or modify
@ -31,193 +31,197 @@ Dis_slots dis_slots;
Prob_prices prob_prices;
bool Mf_read_block( struct Matchfinder * const matchfinder )
bool Mf_read_block( struct Matchfinder * const mf )
{
const int size = matchfinder->buffer_size - matchfinder->stream_pos;
const int rd = readblock( matchfinder->infd_, matchfinder->buffer + matchfinder->stream_pos, size );
matchfinder->stream_pos += rd;
if( rd < size ) matchfinder->at_stream_end = true;
return ( rd == size || !errno );
if( !mf->at_stream_end && mf->stream_pos < mf->buffer_size )
{
const int size = mf->buffer_size - mf->stream_pos;
const int rd = readblock( mf->infd, mf->buffer + mf->stream_pos, size );
mf->stream_pos += rd;
if( rd != size && errno )
{ show_error( "Read error", errno, false ); cleanup_and_fail( 1 ); }
mf->at_stream_end = ( rd < size );
}
return mf->pos < mf->stream_pos;
}
void Mf_init( struct Matchfinder * const matchfinder,
const int dict_size, const int len_limit, const int infd )
void Mf_init( struct Matchfinder * const mf,
const int dict_size, const int len_limit, const int ifd )
{
matchfinder->partial_data_pos = 0;
matchfinder->pos = 0;
matchfinder->cyclic_pos = 0;
matchfinder->stream_pos = 0;
matchfinder->infd_ = infd;
matchfinder->match_len_limit_ = len_limit;
matchfinder->prev_positions =
(int32_t *)malloc( mf_num_prev_positions * sizeof (int32_t) );
if( !matchfinder->prev_positions )
const int buffer_size_limit = ( 2 * dict_size ) + before_size + after_size;
int i;
mf->partial_data_pos = 0;
mf->prev_positions =
(int32_t *)malloc( num_prev_positions * sizeof (int32_t) );
if( !mf->prev_positions )
{
show_error( "not enough memory. Try a smaller dictionary size", 0, false );
show_error( "Not enough memory. Try a smaller dictionary size.", 0, false );
cleanup_and_fail( 1 );
}
matchfinder->at_stream_end = false;
mf->pos = 0;
mf->cyclic_pos = 0;
mf->stream_pos = 0;
mf->match_len_limit_ = len_limit;
mf->cycles = ( len_limit < max_match_len ) ? 16 + ( len_limit / 2 ) : 256;
mf->infd = ifd;
mf->at_stream_end = false;
const int buffer_size_limit = ( 2 * dict_size ) + mf_before_size + mf_after_size;
matchfinder->buffer_size = max( 65536, dict_size );
matchfinder->buffer = (uint8_t *)malloc( matchfinder->buffer_size );
if( !matchfinder->buffer )
mf->buffer_size = max( 65536, dict_size );
mf->buffer = (uint8_t *)malloc( mf->buffer_size );
if( !mf->buffer )
{
show_error( "not enough memory. Try a smaller dictionary size", 0, false );
show_error( "Not enough memory. Try a smaller dictionary size.", 0, false );
cleanup_and_fail( 1 );
}
if( !Mf_read_block( matchfinder ) )
{ show_error( "read error", errno, false ); cleanup_and_fail( 1 ); }
if( !matchfinder->at_stream_end && matchfinder->buffer_size < buffer_size_limit )
if( Mf_read_block( mf ) && !mf->at_stream_end &&
mf->buffer_size < buffer_size_limit )
{
matchfinder->buffer_size = buffer_size_limit;
matchfinder->buffer =
(uint8_t *)realloc( matchfinder->buffer, matchfinder->buffer_size );
if( !matchfinder->buffer )
mf->buffer_size = buffer_size_limit;
mf->buffer = (uint8_t *)realloc( mf->buffer, mf->buffer_size );
if( !mf->buffer )
{
show_error( "not enough memory. Try a smaller dictionary size", 0, false );
show_error( "Not enough memory. Try a smaller dictionary size.", 0, false );
cleanup_and_fail( 1 );
}
if( !Mf_read_block( matchfinder ) )
{ show_error( "read error", errno, false ); cleanup_and_fail( 1 ); }
Mf_read_block( mf );
}
if( matchfinder->at_stream_end && matchfinder->stream_pos < dict_size )
matchfinder->dictionary_size_ = max( min_dictionary_size, matchfinder->stream_pos );
else matchfinder->dictionary_size_ = dict_size;
matchfinder->pos_limit = matchfinder->buffer_size;
if( !matchfinder->at_stream_end ) matchfinder->pos_limit -= mf_after_size;
matchfinder->prev_pos_tree =
(int32_t *)malloc( 2 * matchfinder->dictionary_size_ * sizeof (int32_t) );
if( !matchfinder->prev_pos_tree )
if( mf->at_stream_end && mf->stream_pos < dict_size )
mf->dictionary_size_ = max( min_dictionary_size, mf->stream_pos );
else mf->dictionary_size_ = dict_size;
mf->pos_limit = mf->buffer_size;
if( !mf->at_stream_end ) mf->pos_limit -= after_size;
mf->prev_pos_tree =
(int32_t *)malloc( 2 * mf->dictionary_size_ * sizeof (int32_t) );
if( !mf->prev_pos_tree )
{
show_error( "not enough memory. Try a smaller dictionary size", 0, false );
show_error( "Not enough memory. Try a smaller dictionary size.", 0, false );
cleanup_and_fail( 1 );
}
for( int i = 0; i < mf_num_prev_positions; ++i )
matchfinder->prev_positions[i] = -1;
for( i = 0; i < num_prev_positions; ++i ) mf->prev_positions[i] = -1;
}
bool Mf_reset( struct Matchfinder * const matchfinder )
void Mf_reset( struct Matchfinder * const mf )
{
const int size = matchfinder->stream_pos - matchfinder->pos;
if( size > 0 ) memmove( matchfinder->buffer, matchfinder->buffer + matchfinder->pos, size );
matchfinder->partial_data_pos = 0;
matchfinder->stream_pos -= matchfinder->pos;
matchfinder->pos = 0;
matchfinder->cyclic_pos = 0;
for( int i = 0; i < mf_num_prev_positions; ++i )
matchfinder->prev_positions[i] = -1;
return ( matchfinder->at_stream_end || Mf_read_block( matchfinder ) );
int i;
const int size = mf->stream_pos - mf->pos;
if( size > 0 ) memmove( mf->buffer, mf->buffer + mf->pos, size );
mf->partial_data_pos = 0;
mf->stream_pos -= mf->pos;
mf->pos = 0;
mf->cyclic_pos = 0;
for( i = 0; i < num_prev_positions; ++i ) mf->prev_positions[i] = -1;
Mf_read_block( mf );
}
bool Mf_move_pos( struct Matchfinder * const matchfinder )
void Mf_move_pos( struct Matchfinder * const mf )
{
if( ++matchfinder->cyclic_pos >= matchfinder->dictionary_size_ )
matchfinder->cyclic_pos = 0;
if( ++matchfinder->pos >= matchfinder->pos_limit )
if( ++mf->cyclic_pos >= mf->dictionary_size_ ) mf->cyclic_pos = 0;
if( ++mf->pos >= mf->pos_limit )
{
if( matchfinder->pos > matchfinder->stream_pos )
{ matchfinder->pos = matchfinder->stream_pos; return false; }
if( !matchfinder->at_stream_end )
if( mf->pos > mf->stream_pos )
internal_error( "pos > stream_pos in Mf_move_pos" );
if( !mf->at_stream_end )
{
const int offset = matchfinder->pos - matchfinder->dictionary_size_ - mf_before_size;
const int size = matchfinder->stream_pos - offset;
memmove( matchfinder->buffer, matchfinder->buffer + offset, size );
matchfinder->partial_data_pos += offset;
matchfinder->pos -= offset;
matchfinder->stream_pos -= offset;
for( int i = 0; i < mf_num_prev_positions; ++i )
if( matchfinder->prev_positions[i] >= 0 ) matchfinder->prev_positions[i] -= offset;
for( int i = 0; i < 2 * matchfinder->dictionary_size_; ++i )
if( matchfinder->prev_pos_tree[i] >= 0 ) matchfinder->prev_pos_tree[i] -= offset;
return Mf_read_block( matchfinder );
int i;
const int offset = mf->pos - mf->dictionary_size_ - before_size;
const int size = mf->stream_pos - offset;
memmove( mf->buffer, mf->buffer + offset, size );
mf->partial_data_pos += offset;
mf->pos -= offset;
mf->stream_pos -= offset;
for( i = 0; i < num_prev_positions; ++i )
if( mf->prev_positions[i] >= 0 ) mf->prev_positions[i] -= offset;
for( i = 0; i < 2 * mf->dictionary_size_; ++i )
if( mf->prev_pos_tree[i] >= 0 ) mf->prev_pos_tree[i] -= offset;
Mf_read_block( mf );
}
}
return true;
}
int Mf_longest_match_len( struct Matchfinder * const matchfinder,
int * const distances )
int Mf_longest_match_len( struct Matchfinder * const mf, int * const distances )
{
int len_limit = matchfinder->match_len_limit_;
if( len_limit > Mf_available_bytes( matchfinder ) )
int32_t * ptr0 = mf->prev_pos_tree + ( mf->cyclic_pos << 1 );
int32_t * ptr1 = ptr0 + 1;
int32_t * newptr;
const uint8_t * newdata;
int len = 0, len0 = 0, len1 = 0;
int maxlen = min_match_len - 1;
const int min_pos = (mf->pos >= mf->dictionary_size_) ?
(mf->pos - mf->dictionary_size_ + 1) : 0;
const uint8_t * const data = mf->buffer + mf->pos;
int count, delta, key2, key3, key4, newpos, tmp;
int len_limit = mf->match_len_limit_;
if( len_limit > Mf_available_bytes( mf ) )
{
len_limit = Mf_available_bytes( matchfinder );
len_limit = Mf_available_bytes( mf );
if( len_limit < 4 ) return 0;
}
int maxlen = min_match_len - 1;
const int min_pos = (matchfinder->pos >= matchfinder->dictionary_size_) ?
(matchfinder->pos - matchfinder->dictionary_size_ + 1) : 0;
const uint8_t * const data = matchfinder->buffer + matchfinder->pos;
const int key2 = mf_num_prev_positions4 + mf_num_prev_positions3 +
( ( (int)data[0] << 8 ) | data[1] );
const uint32_t tmp = crc32[data[0]] ^ data[1] ^ ( (uint32_t)data[2] << 8 );
const int key3 = mf_num_prev_positions4 +
(int)( tmp & ( mf_num_prev_positions3 - 1 ) );
const int key4 = (int)( ( tmp ^ ( crc32[data[3]] << 5 ) ) &
( mf_num_prev_positions4 - 1 ) );
key2 = num_prev_positions4 + num_prev_positions3 +
( ( (int)data[0] << 8 ) | data[1] );
tmp = crc32[data[0]] ^ data[1] ^ ( (uint32_t)data[2] << 8 );
key3 = num_prev_positions4 + (int)( tmp & ( num_prev_positions3 - 1 ) );
key4 = (int)( ( tmp ^ ( crc32[data[3]] << 5 ) ) &
( num_prev_positions4 - 1 ) );
if( distances )
{
int np = matchfinder->prev_positions[key2];
int np = mf->prev_positions[key2];
if( np >= min_pos )
{ distances[2] = matchfinder->pos - np - 1; maxlen = 2; }
{ distances[2] = mf->pos - np - 1; maxlen = 2; }
else distances[2] = 0x7FFFFFFF;
np = matchfinder->prev_positions[key3];
if( np >= min_pos && matchfinder->buffer[np] == data[0] )
{ distances[3] = matchfinder->pos - np - 1; maxlen = 3; }
np = mf->prev_positions[key3];
if( np >= min_pos && mf->buffer[np] == data[0] )
{ distances[3] = mf->pos - np - 1; maxlen = 3; }
else distances[3] = 0x7FFFFFFF;
distances[4] = 0x7FFFFFFF;
}
matchfinder->prev_positions[key2] = matchfinder->pos;
matchfinder->prev_positions[key3] = matchfinder->pos;
int newpos = matchfinder->prev_positions[key4];
matchfinder->prev_positions[key4] = matchfinder->pos;
mf->prev_positions[key2] = mf->pos;
mf->prev_positions[key3] = mf->pos;
newpos = mf->prev_positions[key4];
mf->prev_positions[key4] = mf->pos;
int idx0 = matchfinder->cyclic_pos << 1;
int idx1 = idx0 + 1;
int len = 0, len0 = 0, len1 = 0;
for( int count = 16 + ( matchfinder->match_len_limit_ / 2 ); ; )
for( count = mf->cycles; ; )
{
if( newpos < min_pos || --count < 0 )
{ matchfinder->prev_pos_tree[idx0] = matchfinder->prev_pos_tree[idx1] = -1; break; }
const uint8_t * const newdata = matchfinder->buffer + newpos;
if( newpos < min_pos || --count < 0 ) { *ptr0 = *ptr1 = -1; break; }
newdata = mf->buffer + newpos;
while( len < len_limit && newdata[len] == data[len] ) ++len;
const int delta = matchfinder->pos - newpos;
delta = mf->pos - newpos;
if( distances ) while( maxlen < len ) distances[++maxlen] = delta - 1;
const int newidx = ( matchfinder->cyclic_pos - delta +
( ( matchfinder->cyclic_pos >= delta ) ? 0 : matchfinder->dictionary_size_ ) ) << 1;
newptr = mf->prev_pos_tree +
( ( mf->cyclic_pos - delta +
( ( mf->cyclic_pos >= delta ) ? 0 : mf->dictionary_size_ ) ) << 1 );
if( len < len_limit )
{
if( newdata[len] < data[len] )
{
matchfinder->prev_pos_tree[idx0] = newpos;
idx0 = newidx + 1;
newpos = matchfinder->prev_pos_tree[idx0];
*ptr0 = newpos;
ptr0 = newptr + 1;
newpos = *ptr0;
len0 = len; if( len1 < len ) len = len1;
}
else
{
matchfinder->prev_pos_tree[idx1] = newpos;
idx1 = newidx;
newpos = matchfinder->prev_pos_tree[idx1];
*ptr1 = newpos;
ptr1 = newptr;
newpos = *ptr1;
len1 = len; if( len0 < len ) len = len0;
}
}
else
{
matchfinder->prev_pos_tree[idx0] = matchfinder->prev_pos_tree[newidx];
matchfinder->prev_pos_tree[idx1] = matchfinder->prev_pos_tree[newidx+1];
*ptr0 = newptr[0];
*ptr1 = newptr[1];
break;
}
}
@ -230,6 +234,20 @@ int Mf_longest_match_len( struct Matchfinder * const matchfinder,
}
void Re_flush_data( struct Range_encoder * const range_encoder )
{
if( range_encoder->pos > 0 )
{
if( range_encoder->outfd >= 0 &&
writeblock( range_encoder->outfd, range_encoder->buffer,
range_encoder->pos ) != range_encoder->pos )
{ show_error( "Write error", errno, false ); cleanup_and_fail( 1 ); }
range_encoder->partial_member_pos += range_encoder->pos;
range_encoder->pos = 0;
}
}
void Lee_encode( struct Len_encoder * const len_encoder,
struct Range_encoder * const range_encoder,
int symbol, const int pos_state )
@ -261,7 +279,8 @@ void Lee_encode( struct Len_encoder * const len_encoder,
void LZe_fill_align_prices( struct LZ_encoder * const encoder )
{
for( int i = 0; i < dis_align_size; ++i )
int i;
for( i = 0; i < dis_align_size; ++i )
encoder->align_prices[i] = price_symbol_reversed( encoder->bm_align, i, dis_align_bits );
encoder->align_price_count = dis_align_size;
}
@ -269,9 +288,22 @@ void LZe_fill_align_prices( struct LZ_encoder * const encoder )
void LZe_fill_distance_prices( struct LZ_encoder * const encoder )
{
for( int dis_state = 0; dis_state < max_dis_states; ++dis_state )
int dis, dis_state;
for( dis = start_dis_model; dis < modeled_distances; ++dis )
{
const int dis_slot = dis_slots[dis];
const int direct_bits = ( dis_slot >> 1 ) - 1;
const int base = ( 2 | ( dis_slot & 1 ) ) << direct_bits;
const int price =
price_symbol_reversed( encoder->bm_dis + base - dis_slot, dis - base, direct_bits );
for( dis_state = 0; dis_state < max_dis_states; ++dis_state )
encoder->dis_prices[dis_state][dis] = price;
}
for( dis_state = 0; dis_state < max_dis_states; ++dis_state )
{
int * const dsp = encoder->dis_slot_prices[dis_state];
int * const dp = encoder->dis_prices[dis_state];
const Bit_model * const bmds = encoder->bm_dis_slot[dis_state];
int slot = 0;
for( ; slot < end_dis_model && slot < encoder->num_dis_slots; ++slot )
@ -280,39 +312,32 @@ void LZe_fill_distance_prices( struct LZ_encoder * const encoder )
dsp[slot] = price_symbol( bmds, slot, dis_slot_bits ) +
(((( slot >> 1 ) - 1 ) - dis_align_bits ) << price_shift );
int * const dp = encoder->dis_prices[dis_state];
int dis = 0;
for( ; dis < start_dis_model; ++dis )
for( dis = 0; dis < start_dis_model; ++dis )
dp[dis] = dsp[dis];
for( ; dis < modeled_distances; ++dis )
{
const int dis_slot = get_slot( dis );
const int direct_bits = ( dis_slot >> 1 ) - 1;
const int base = ( 2 | ( dis_slot & 1 ) ) << direct_bits;
dp[dis] = dsp[dis_slot] +
price_symbol_reversed( encoder->bm_dis + base - dis_slot, dis - base, direct_bits );
}
dp[dis] += dsp[dis_slots[dis]];
}
}
// Return value == number of bytes advanced (ahead).
// trials[0]..trials[retval-1] contain the steps to encode.
// ( trials[0].dis == -1 && trials[0].price == 1 ) means literal.
/* Return value == number of bytes advanced (ahead).
trials[0]..trials[retval-1] contain the steps to encode.
( trials[0].dis == -1 && trials[0].price == 1 ) means literal. */
int LZe_sequence_optimizer( struct LZ_encoder * const encoder,
const int reps[num_rep_distances], const State state )
{
int main_len;
if( encoder->longest_match_found > 0 ) // from previous call
int main_len, i, rep, cur = 0, num_trials;
int replens[num_rep_distances];
int rep_index = 0;
if( encoder->longest_match_found > 0 ) /* from previous call */
{
main_len = encoder->longest_match_found;
encoder->longest_match_found = 0;
}
else main_len = LZe_read_match_distances( encoder );
int replens[num_rep_distances];
int rep_index = 0;
for( int i = 0; i < num_rep_distances; ++i )
for( i = 0; i < num_rep_distances; ++i )
{
replens[i] = Mf_true_match_len( encoder->matchfinder, 0, reps[i] + 1, max_match_len );
if( replens[i] > replens[rep_index] ) rep_index = i;
@ -321,7 +346,7 @@ int LZe_sequence_optimizer( struct LZ_encoder * const encoder,
{
encoder->trials[0].dis = rep_index;
encoder->trials[0].price = replens[rep_index];
if( !LZe_move_pos( encoder, replens[rep_index], true ) ) return 0;
LZe_move_pos( encoder, replens[rep_index], true );
return replens[rep_index];
}
@ -330,18 +355,20 @@ int LZe_sequence_optimizer( struct LZ_encoder * const encoder,
encoder->trials[0].dis = encoder->match_distances[Mf_match_len_limit( encoder->matchfinder )] +
num_rep_distances;
encoder->trials[0].price = main_len;
if( !LZe_move_pos( encoder, main_len, true ) ) return 0;
LZe_move_pos( encoder, main_len, true );
return main_len;
}
{
const int pos_state = Mf_data_position( encoder->matchfinder ) & pos_state_mask;
const int match_price = price1( encoder->bm_match[state][pos_state] );
const int rep_match_price = match_price + price1( encoder->bm_rep[state] );
const uint8_t prev_byte = Mf_peek( encoder->matchfinder, -1 );
const uint8_t cur_byte = Mf_peek( encoder->matchfinder, 0 );
const uint8_t match_byte = Mf_peek( encoder->matchfinder, -reps[0]-1 );
encoder->trials[0].state = state;
for( int i = 0; i < num_rep_distances; ++i ) encoder->trials[0].reps[i] = reps[i];
for( i = 0; i < num_rep_distances; ++i ) encoder->trials[0].reps[i] = reps[i];
encoder->trials[1].dis = -1;
encoder->trials[1].prev_index = 0;
encoder->trials[1].price = price0( encoder->bm_match[state][pos_state] );
@ -350,9 +377,6 @@ int LZe_sequence_optimizer( struct LZ_encoder * const encoder,
else
encoder->trials[1].price += Lie_price_matched( &encoder->literal_encoder, prev_byte, cur_byte, match_byte );
const int match_price = price1( encoder->bm_match[state][pos_state] );
const int rep_match_price = match_price + price1( encoder->bm_rep[state] );
if( match_byte == cur_byte )
Tr_update( &encoder->trials[1], 0, 0, rep_match_price + LZe_price_rep_len1( encoder, state, pos_state ) );
@ -360,7 +384,7 @@ int LZe_sequence_optimizer( struct LZ_encoder * const encoder,
{
encoder->trials[0].dis = encoder->trials[1].dis;
encoder->trials[0].price = 1;
if( !Mf_move_pos( encoder->matchfinder ) ) return 0;
Mf_move_pos( encoder->matchfinder );
return 1;
}
@ -371,7 +395,7 @@ int LZe_sequence_optimizer( struct LZ_encoder * const encoder,
{
main_len = replens[rep_index];
for( ; len <= main_len; ++len )
encoder->trials[len].price = lze_infinite_price;
encoder->trials[len].price = infinite_price;
}
else for( ; len <= main_len; ++len )
{
@ -382,28 +406,33 @@ int LZe_sequence_optimizer( struct LZ_encoder * const encoder,
}
}
for( int rep = 0; rep < num_rep_distances; ++rep )
for( rep = 0; rep < num_rep_distances; ++rep )
{
const int price = rep_match_price +
LZe_price_rep( encoder, rep, state, pos_state );
for( int len = min_match_len; len <= replens[rep]; ++len )
int len;
for( len = min_match_len; len <= replens[rep]; ++len )
Tr_update( &encoder->trials[len], rep, 0, price +
Lee_price( &encoder->rep_match_len_encoder, len, pos_state ) );
}
}
int cur = 0;
int num_trials = main_len;
if( !Mf_move_pos( encoder->matchfinder ) ) return 0;
num_trials = main_len;
Mf_move_pos( encoder->matchfinder );
while( true )
{
if( ++cur >= num_trials ) // no more initialized trials
struct Trial *cur_trial, *next_trial;
int newlen, pos_state, prev_index, len_limit;
int next_price, match_price, rep_match_price;
uint8_t prev_byte, cur_byte, match_byte;
if( ++cur >= num_trials ) /* no more initialized trials */
{
LZe_backward( encoder, cur );
return cur;
}
const int newlen = LZe_read_match_distances( encoder );
newlen = LZe_read_match_distances( encoder );
if( newlen >= Mf_match_len_limit( encoder->matchfinder ) )
{
encoder->longest_match_found = newlen;
@ -411,12 +440,12 @@ int LZe_sequence_optimizer( struct LZ_encoder * const encoder,
return cur;
}
struct Trial * const cur_trial = &encoder->trials[cur];
const int prev_index = cur_trial->prev_index;
cur_trial = &encoder->trials[cur];
prev_index = cur_trial->prev_index;
cur_trial->state = encoder->trials[prev_index].state;
for( int i = 0; i < num_rep_distances; ++i )
for( i = 0; i < num_rep_distances; ++i )
cur_trial->reps[i] = encoder->trials[prev_index].reps[i];
if( prev_index == cur - 1 )
{
@ -430,35 +459,35 @@ int LZe_sequence_optimizer( struct LZ_encoder * const encoder,
LZe_mtf_reps( cur_trial->dis, cur_trial->reps );
}
const int pos_state = Mf_data_position( encoder->matchfinder ) & pos_state_mask;
const uint8_t prev_byte = Mf_peek( encoder->matchfinder, -1 );
const uint8_t cur_byte = Mf_peek( encoder->matchfinder, 0 );
const uint8_t match_byte = Mf_peek( encoder->matchfinder, -cur_trial->reps[0]-1 );
pos_state = Mf_data_position( encoder->matchfinder ) & pos_state_mask;
prev_byte = Mf_peek( encoder->matchfinder, -1 );
cur_byte = Mf_peek( encoder->matchfinder, 0 );
match_byte = Mf_peek( encoder->matchfinder, -cur_trial->reps[0]-1 );
int next_price = cur_trial->price + price0( encoder->bm_match[cur_trial->state][pos_state] );
next_price = cur_trial->price + price0( encoder->bm_match[cur_trial->state][pos_state] );
if( St_is_char( cur_trial->state ) )
next_price += Lie_price_symbol( &encoder->literal_encoder, prev_byte, cur_byte );
else
next_price += Lie_price_matched( &encoder->literal_encoder, prev_byte, cur_byte, match_byte );
if( !Mf_move_pos( encoder->matchfinder ) ) return 0;
Mf_move_pos( encoder->matchfinder );
struct Trial * const next_trial = &encoder->trials[cur+1];
next_trial = &encoder->trials[cur+1];
Tr_update( next_trial, -1, cur, next_price );
const int match_price = cur_trial->price + price1( encoder->bm_match[cur_trial->state][pos_state] );
const int rep_match_price = match_price + price1( encoder->bm_rep[cur_trial->state] );
match_price = cur_trial->price + price1( encoder->bm_match[cur_trial->state][pos_state] );
rep_match_price = match_price + price1( encoder->bm_rep[cur_trial->state] );
if( match_byte == cur_byte && next_trial->dis != 0 )
Tr_update( next_trial, 0, cur, rep_match_price +
LZe_price_rep_len1( encoder, cur_trial->state, pos_state ) );
const int len_limit = min( min( max_num_trials - 1 - cur,
len_limit = min( min( max_num_trials - 1 - cur,
Mf_available_bytes( encoder->matchfinder ) ),
Mf_match_len_limit( encoder->matchfinder ) );
Mf_match_len_limit( encoder->matchfinder ) );
if( len_limit < min_match_len ) continue;
for( int rep = 0; rep < num_rep_distances; ++rep )
for( rep = 0; rep < num_rep_distances; ++rep )
{
const int dis = cur_trial->reps[rep] + 1;
int len = 0;
@ -469,7 +498,7 @@ int LZe_sequence_optimizer( struct LZ_encoder * const encoder,
const int price = rep_match_price +
LZe_price_rep( encoder, rep, cur_trial->state, pos_state );
while( num_trials < cur + len )
encoder->trials[++num_trials].price = lze_infinite_price;
encoder->trials[++num_trials].price = infinite_price;
for( ; len >= min_match_len; --len )
Tr_update( &encoder->trials[cur+len], rep, cur, price +
Lee_price( &encoder->rep_match_len_encoder, len, pos_state ) );
@ -483,10 +512,11 @@ int LZe_sequence_optimizer( struct LZ_encoder * const encoder,
{
const int normal_match_price = match_price +
price0( encoder->bm_rep[cur_trial->state] );
int len;
while( num_trials < cur + newlen )
encoder->trials[++num_trials].price = lze_infinite_price;
encoder->trials[++num_trials].price = infinite_price;
for( int len = min_match_len; len <= newlen; ++len )
for( len = min_match_len; len <= newlen; ++len )
Tr_update( &encoder->trials[cur+len], encoder->match_distances[len] + num_rep_distances,
cur, normal_match_price +
LZe_price_pair( encoder, encoder->match_distances[len], len, pos_state ) );
@ -495,19 +525,20 @@ int LZe_sequence_optimizer( struct LZ_encoder * const encoder,
}
// End Of Stream mark => (dis == 0xFFFFFFFFU, len == min_match_len)
/* End Of Stream mark => (dis == 0xFFFFFFFFU, len == min_match_len) */
void LZe_full_flush( struct LZ_encoder * const encoder, const State state )
{
int i;
const int pos_state = Mf_data_position( encoder->matchfinder ) & pos_state_mask;
File_trailer trailer;
Re_encode_bit( &encoder->range_encoder, &encoder->bm_match[state][pos_state], 1 );
Re_encode_bit( &encoder->range_encoder, &encoder->bm_rep[state], 0 );
LZe_encode_pair( encoder, 0xFFFFFFFFU, min_match_len, pos_state );
Re_flush( &encoder->range_encoder );
File_trailer trailer;
Ft_set_data_crc( trailer, LZe_crc( encoder ) );
Ft_set_data_size( trailer, Mf_data_position( encoder->matchfinder ) );
Ft_set_member_size( trailer, LZe_member_position( encoder ) + Ft_size );
for( int i = 0; i < Ft_size; ++i )
for( i = 0; i < Ft_size; ++i )
Re_put_byte( &encoder->range_encoder, trailer[i] );
Re_flush_data( &encoder->range_encoder );
}
@ -516,12 +547,13 @@ void LZe_full_flush( struct LZ_encoder * const encoder, const State state )
void LZe_init( struct LZ_encoder * const encoder, struct Matchfinder * const mf,
const File_header header, const int outfd )
{
int i, j;
encoder->longest_match_found = 0;
encoder->crc_ = 0xFFFFFFFFU;
for( int i = 0; i < St_states; ++i )
for( i = 0; i < states; ++i )
{
for( int j = 0; j < pos_states; ++j )
for( j = 0; j < pos_states; ++j )
{
Bm_init( &encoder->bm_match[i][j] );
Bm_init( &encoder->bm_len[i][j] );
@ -531,12 +563,12 @@ void LZe_init( struct LZ_encoder * const encoder, struct Matchfinder * const mf,
Bm_init( &encoder->bm_rep1[i] );
Bm_init( &encoder->bm_rep2[i] );
}
for( int i = 0; i < max_dis_states; ++i )
for( int j = 0; j < 1<<dis_slot_bits; ++j )
for( i = 0; i < max_dis_states; ++i )
for( j = 0; j < 1<<dis_slot_bits; ++j )
Bm_init( &encoder->bm_dis_slot[i][j] );
for( int i = 0; i < modeled_distances-end_dis_model; ++i )
for( i = 0; i < modeled_distances-end_dis_model+1; ++i )
Bm_init( &encoder->bm_dis[i] );
for( int i = 0; i < dis_align_size; ++i )
for( i = 0; i < dis_align_size; ++i )
Bm_init( &encoder->bm_align[i] );
encoder->matchfinder = mf;
@ -544,36 +576,41 @@ void LZe_init( struct LZ_encoder * const encoder, struct Matchfinder * const mf,
Lee_init( &encoder->len_encoder, Mf_match_len_limit( encoder->matchfinder ) ),
Lee_init( &encoder->rep_match_len_encoder, Mf_match_len_limit( encoder->matchfinder ) ),
Lie_init( &encoder->literal_encoder );
encoder->num_dis_slots = 2 * Fh_real_bits( Mf_dictionary_size( encoder->matchfinder ) - 1 );
encoder->num_dis_slots = 2 * real_bits( Mf_dictionary_size( encoder->matchfinder ) - 1 );
LZe_fill_align_prices( encoder );
for( int i = 0; i < Fh_size; ++i )
for( i = 0; i < Fh_size; ++i )
Re_put_byte( &encoder->range_encoder, header[i] );
}
bool LZe_encode_member( struct LZ_encoder * const encoder, const long long member_size )
bool LZe_encode_member( struct LZ_encoder * const encoder,
const long long member_size )
{
if( LZe_member_position( encoder ) != Fh_size )
return false; // can be called only once
const long long member_size_limit =
member_size - Ft_size - lze_max_marker_size;
member_size - Ft_size - max_marker_size;
const int fill_count =
( Mf_match_len_limit( encoder->matchfinder ) > 12 ) ? 512 : 2048;
int fill_counter = 0;
int ahead;
int i;
int rep_distances[num_rep_distances];
State state = 0;
for( int i = 0; i < num_rep_distances; ++i ) rep_distances[i] = 0;
for( i = 0; i < num_rep_distances; ++i ) rep_distances[i] = 0;
// encode first byte
if( Mf_data_position( encoder->matchfinder ) == 0 &&
!Mf_finished( encoder->matchfinder ) )
if( Mf_data_position( encoder->matchfinder ) != 0 ||
LZe_member_position( encoder ) != Fh_size )
return false; /* can be called only once */
if( !Mf_finished( encoder->matchfinder ) ) /* encode first byte */
{
Re_encode_bit( &encoder->range_encoder, &encoder->bm_match[state][0], 0 );
const uint8_t prev_byte = 0;
const uint8_t cur_byte = Mf_peek( encoder->matchfinder, 0 );
Re_encode_bit( &encoder->range_encoder, &encoder->bm_match[state][0], 0 );
Lie_encode( &encoder->literal_encoder, &encoder->range_encoder, prev_byte, cur_byte );
CRC32_update_byte( &encoder->crc_, cur_byte );
if( !LZe_move_pos( encoder, 1, false ) ) return false;
LZe_move_pos( encoder, 1, false );
}
while( true )
@ -581,13 +618,13 @@ bool LZe_encode_member( struct LZ_encoder * const encoder, const long long membe
if( Mf_finished( encoder->matchfinder ) )
{ LZe_full_flush( encoder, state ); return true; }
if( fill_counter <= 0 )
{ LZe_fill_distance_prices( encoder ); fill_counter = 512; }
{ LZe_fill_distance_prices( encoder ); fill_counter = fill_count; }
int ahead = LZe_sequence_optimizer( encoder, rep_distances, state );
ahead = LZe_sequence_optimizer( encoder, rep_distances, state );
if( ahead <= 0 ) return false;
fill_counter -= ahead;
for( int i = 0; ; )
for( i = 0; ; )
{
const int pos_state = ( Mf_data_position( encoder->matchfinder ) - ahead ) & pos_state_mask;
const int dis = encoder->trials[i].dis;
@ -595,7 +632,7 @@ bool LZe_encode_member( struct LZ_encoder * const encoder, const long long membe
bool bit = ( dis < 0 && len == 1 );
Re_encode_bit( &encoder->range_encoder, &encoder->bm_match[state][pos_state], !bit );
if( bit ) // literal byte
if( bit ) /* literal byte */
{
const uint8_t prev_byte = Mf_peek( encoder->matchfinder, -ahead-1 );
const uint8_t cur_byte = Mf_peek( encoder->matchfinder, -ahead );
@ -605,11 +642,11 @@ bool LZe_encode_member( struct LZ_encoder * const encoder, const long long membe
else
{
const uint8_t match_byte = Mf_peek( encoder->matchfinder, -ahead-rep_distances[0]-1 );
Lie_encode_matched( &encoder->literal_encoder, &encoder->range_encoder, prev_byte, match_byte, cur_byte );
Lie_encode_matched( &encoder->literal_encoder, &encoder->range_encoder, prev_byte, cur_byte, match_byte );
}
St_set_char( &state );
}
else // match or repeated match
else /* match or repeated match */
{
CRC32_update_buf( &encoder->crc_, Mf_ptr_to_current_pos( encoder->matchfinder ) - ahead, len );
LZe_mtf_reps( dis, rep_distances );

292
encoder.h
View file

@ -1,4 +1,4 @@
/* Clzip - A data compressor based on the LZMA algorithm
/* Clzip - Data compressor based on the LZMA algorithm
Copyright (C) 2010 Antonio Diaz Diaz.
This program is free software: you can redistribute it and/or modify
@ -24,8 +24,9 @@ extern Dis_slots dis_slots;
static inline void Dis_slots_init()
{
for( int slot = 0; slot < 4; ++slot ) dis_slots[slot] = slot;
for( int i = 4, size = 2, slot = 4; slot < 24; slot += 2 )
int i, size, slot;
for( slot = 0; slot < 4; ++slot ) dis_slots[slot] = slot;
for( i = 4, size = 2, slot = 4; slot < 24; slot += 2 )
{
memset( &dis_slots[i], slot, size );
memset( &dis_slots[i+size], slot + 1, size );
@ -49,13 +50,13 @@ extern Prob_prices prob_prices;
static inline void Prob_prices_init()
{
const int num_bits = ( bit_model_total_bits - 2 );
for( int i = num_bits - 1; i >= 0; --i )
int i, j = 1, end = 2;
prob_prices[0] = bit_model_total_bits << price_shift;
for( i = num_bits - 1; i >= 0; --i, end <<= 1 )
{
int start = 1 << ( num_bits - i - 1 );
int end = 1 << ( num_bits - i);
for( int j = start; j < end; ++j )
prob_prices[j] = (i << price_shift) +
( ((end - j) << price_shift) >> (num_bits - i - 1) );
for( ; j < end; ++j )
prob_prices[j] = ( i << price_shift ) +
( ((end - j) << price_shift) >> (num_bits - i - 1) );
}
}
@ -73,10 +74,11 @@ static inline int price_bit( const Bit_model bm, const int bit )
{ if( bit ) return price1( bm ); else return price0( bm ); }
static inline int price_symbol( const Bit_model bm[], int symbol, const int num_bits )
static inline int price_symbol( const Bit_model bm[], int symbol,
const int num_bits )
{
symbol |= ( 1 << num_bits );
int price = 0;
symbol |= ( 1 << num_bits );
while( symbol > 1 )
{
const int bit = symbol & 1;
@ -92,7 +94,8 @@ static inline int price_symbol_reversed( const Bit_model bm[], int symbol,
{
int price = 0;
int model = 1;
for( int i = num_bits; i > 0; --i )
int i;
for( i = num_bits; i > 0; --i )
{
const int bit = symbol & 1;
symbol >>= 1;
@ -108,8 +111,9 @@ static inline int price_matched( const Bit_model bm[], const int symbol,
{
int price = 0;
int model = 1;
int i;
for( int i = 7; i >= 0; --i )
for( i = 7; i >= 0; --i )
{
const int match_bit = ( match_byte >> i ) & 1;
int bit = ( symbol >> i ) & 1;
@ -130,86 +134,87 @@ static inline int price_matched( const Bit_model bm[], const int symbol,
}
enum { // bytes to keep in buffer before dictionary
mf_before_size = max_num_trials + 1,
// bytes to keep in buffer after pos
mf_after_size = max_match_len,
mf_num_prev_positions4 = 1 << 20,
mf_num_prev_positions3 = 1 << 18,
mf_num_prev_positions2 = 1 << 16,
mf_num_prev_positions = mf_num_prev_positions4 + mf_num_prev_positions3 +
mf_num_prev_positions2 };
enum { /* bytes to keep in buffer before dictionary */
before_size = max_num_trials + 1,
/* bytes to keep in buffer after pos */
after_size = max_match_len,
num_prev_positions4 = 1 << 20,
num_prev_positions3 = 1 << 18,
num_prev_positions2 = 1 << 16,
num_prev_positions = num_prev_positions4 + num_prev_positions3 +
num_prev_positions2 };
struct Matchfinder
{
long long partial_data_pos;
int dictionary_size_; // bytes to keep in buffer before pos
int buffer_size;
uint8_t * buffer;
int pos;
int cyclic_pos;
int stream_pos; // first byte not yet read from file
int pos_limit; // when reached, a new block must be read
int infd_; // input file descriptor
int match_len_limit_;
int32_t * prev_positions; // last seen position of key
uint8_t * buffer; /* input buffer */
int32_t * prev_positions; /* last seen position of key */
int32_t * prev_pos_tree;
bool at_stream_end; // stream_pos shows real end of file
int dictionary_size_; /* bytes to keep in buffer before pos */
int buffer_size;
int pos; /* current pos in buffer */
int cyclic_pos; /* current pos in dictionary */
int stream_pos; /* first byte not yet read from file */
int pos_limit; /* when reached, a new block must be read */
int match_len_limit_;
int cycles;
int infd; /* input file descriptor */
bool at_stream_end; /* stream_pos shows real end of file */
};
bool Mf_read_block( struct Matchfinder * const matchfinder );
bool Mf_read_block( struct Matchfinder * const mf );
void Mf_init( struct Matchfinder * const matchfinder,
const int dict_size, const int len_limit, const int infd );
void Mf_init( struct Matchfinder * const mf,
const int dict_size, const int len_limit, const int ifd );
static inline void Mf_free( struct Matchfinder * const matchfinder )
static inline void Mf_free( struct Matchfinder * const mf )
{
free( matchfinder->prev_pos_tree ); matchfinder->prev_pos_tree = 0;
free( matchfinder->prev_positions ); matchfinder->prev_positions = 0;
free( matchfinder->buffer ); matchfinder->buffer = 0;
free( mf->prev_pos_tree ); mf->prev_pos_tree = 0;
free( mf->prev_positions ); mf->prev_positions = 0;
free( mf->buffer ); mf->buffer = 0;
}
static inline uint8_t Mf_peek( struct Matchfinder * const matchfinder, const int i )
{ return matchfinder->buffer[matchfinder->pos+i]; }
static inline int Mf_available_bytes( struct Matchfinder * const matchfinder )
{ return matchfinder->stream_pos - matchfinder->pos; }
static inline long long Mf_data_position( struct Matchfinder * const matchfinder )
{ return matchfinder->partial_data_pos + matchfinder->pos; }
static inline int Mf_dictionary_size( struct Matchfinder * const matchfinder )
{ return matchfinder->dictionary_size_; }
static inline bool Mf_finished( struct Matchfinder * const matchfinder )
{ return matchfinder->at_stream_end && matchfinder->pos >= matchfinder->stream_pos; }
static inline int Mf_match_len_limit( struct Matchfinder * const matchfinder )
{ return matchfinder->match_len_limit_; }
static inline const uint8_t * Mf_ptr_to_current_pos( struct Matchfinder * const matchfinder )
{ return matchfinder->buffer + matchfinder->pos; }
static inline uint8_t Mf_peek( struct Matchfinder * const mf, const int i )
{ return mf->buffer[mf->pos+i]; }
static inline int Mf_available_bytes( struct Matchfinder * const mf )
{ return mf->stream_pos - mf->pos; }
static inline long long Mf_data_position( struct Matchfinder * const mf )
{ return mf->partial_data_pos + mf->pos; }
static inline int Mf_dictionary_size( struct Matchfinder * const mf )
{ return mf->dictionary_size_; }
static inline bool Mf_finished( struct Matchfinder * const mf )
{ return mf->at_stream_end && mf->pos >= mf->stream_pos; }
static inline int Mf_match_len_limit( struct Matchfinder * const mf )
{ return mf->match_len_limit_; }
static inline const uint8_t * Mf_ptr_to_current_pos( struct Matchfinder * const mf )
{ return mf->buffer + mf->pos; }
static inline bool Mf_dec_pos( struct Matchfinder * const matchfinder,
static inline bool Mf_dec_pos( struct Matchfinder * const mf,
const int ahead )
{
if( ahead < 0 || matchfinder->pos < ahead ) return false;
matchfinder->pos -= ahead;
matchfinder->cyclic_pos -= ahead;
if( matchfinder->cyclic_pos < 0 )
matchfinder->cyclic_pos += matchfinder->dictionary_size_;
if( ahead < 0 || mf->pos < ahead ) return false;
mf->pos -= ahead;
mf->cyclic_pos -= ahead;
if( mf->cyclic_pos < 0 ) mf->cyclic_pos += mf->dictionary_size_;
return true;
}
static inline int Mf_true_match_len( struct Matchfinder * const matchfinder,
const int index, const int distance, int len_limit )
static inline int Mf_true_match_len( struct Matchfinder * const mf,
const int index, const int distance,
int len_limit )
{
if( index + len_limit > Mf_available_bytes( matchfinder ) )
len_limit = Mf_available_bytes( matchfinder ) - index;
const uint8_t * const data = matchfinder->buffer + matchfinder->pos + index - distance;
const uint8_t * const data = mf->buffer + mf->pos + index - distance;
int i = 0;
if( index + len_limit > Mf_available_bytes( mf ) )
len_limit = Mf_available_bytes( mf ) - index;
while( i < len_limit && data[i] == data[i+distance] ) ++i;
return i;
}
bool Mf_reset( struct Matchfinder * const matchfinder );
bool Mf_move_pos( struct Matchfinder * const matchfinder );
int Mf_longest_match_len( struct Matchfinder * const matchfinder,
int * const distances );
void Mf_reset( struct Matchfinder * const mf );
void Mf_move_pos( struct Matchfinder * const mf );
int Mf_longest_match_len( struct Matchfinder * const mf, int * const distances );
enum { re_buffer_size = 65536 };
@ -218,28 +223,15 @@ struct Range_encoder
{
uint64_t low;
long long partial_member_pos;
uint8_t * buffer;
int pos;
uint8_t * buffer; /* output buffer */
int pos; /* current pos in buffer */
uint32_t range;
int ff_count;
int outfd_; // output file descriptor
int outfd; /* output file descriptor */
uint8_t cache;
};
static inline void Re_flush_data( struct Range_encoder * const range_encoder )
{
if( range_encoder->pos > 0 )
{
if( range_encoder->outfd_ >= 0 )
{
const int wr = writeblock( range_encoder->outfd_, range_encoder->buffer, range_encoder->pos );
if( wr != range_encoder->pos )
{ show_error( "write error", errno, false ); cleanup_and_fail( 1 ); }
}
range_encoder->partial_member_pos += range_encoder->pos;
range_encoder->pos = 0;
}
}
void Re_flush_data( struct Range_encoder * const range_encoder );
static inline void Re_put_byte( struct Range_encoder * const range_encoder,
const uint8_t b )
@ -263,36 +255,37 @@ static inline void Re_shift_low( struct Range_encoder * const range_encoder )
}
static inline void Re_init( struct Range_encoder * const range_encoder,
const int outfd )
const int ofd )
{
range_encoder->low = 0;
range_encoder->partial_member_pos = 0;
range_encoder->buffer = (uint8_t *)malloc( re_buffer_size );
if( !range_encoder->buffer )
{
show_error( "not enough memory. Try a smaller dictionary size", 0, false );
show_error( "Not enough memory. Try a smaller dictionary size.", 0, false );
cleanup_and_fail( 1 );
}
range_encoder->pos = 0;
range_encoder->range = 0xFFFFFFFFU;
range_encoder->ff_count = 0;
range_encoder->outfd_ = outfd;
range_encoder->outfd = ofd;
range_encoder->cache = 0;
}
static inline void Re_free( struct Range_encoder * const range_encoder )
{ free( range_encoder->buffer ); range_encoder->buffer = 0; }
static inline void Re_flush( struct Range_encoder * const range_encoder )
{ for( int i = 0; i < 5; ++i ) Re_shift_low( range_encoder ); }
static inline long long Re_member_position( struct Range_encoder * const range_encoder )
{ return range_encoder->partial_member_pos + range_encoder->pos + range_encoder->ff_count; }
static inline void Re_flush( struct Range_encoder * const range_encoder )
{ int i; for( i = 0; i < 5; ++i ) Re_shift_low( range_encoder ); }
static inline void Re_encode( struct Range_encoder * const range_encoder,
const int symbol, const int num_bits )
{
for( int i = num_bits - 1; i >= 0; --i )
int i;
for( i = num_bits - 1; i >= 0; --i )
{
range_encoder->range >>= 1;
if( (symbol >> i) & 1 ) range_encoder->low += range_encoder->range;
@ -325,7 +318,8 @@ static inline void Re_encode_tree( struct Range_encoder * const range_encoder,
{
int mask = ( 1 << ( num_bits - 1 ) );
int model = 1;
for( int i = num_bits; i > 0; --i, mask >>= 1 )
int i;
for( i = num_bits; i > 0; --i, mask >>= 1 )
{
const int bit = ( symbol & mask );
Re_encode_bit( range_encoder, &bm[model], bit );
@ -338,7 +332,8 @@ static inline void Re_encode_tree_reversed( struct Range_encoder * const range_e
Bit_model bm[], int symbol, const int num_bits )
{
int model = 1;
for( int i = num_bits; i > 0; --i )
int i;
for( i = num_bits; i > 0; --i )
{
const int bit = symbol & 1;
Re_encode_bit( range_encoder, &bm[model], bit );
@ -351,7 +346,8 @@ static inline void Re_encode_matched( struct Range_encoder * const range_encoder
Bit_model bm[], int symbol, int match_byte )
{
int model = 1;
for( int i = 7; i >= 0; --i )
int i;
for( i = 7; i >= 0; --i )
{
const int match_bit = ( match_byte >> i ) & 1;
int bit = ( symbol >> i ) & 1;
@ -405,18 +401,19 @@ static inline void Lee_update_prices( struct Len_encoder * const len_encoder,
static inline void Lee_init( struct Len_encoder * const len_encoder,
const int len_limit )
{
int i, j;
Bm_init( &len_encoder->choice1 );
Bm_init( &len_encoder->choice2 );
for( int i = 0; i < pos_states; ++i )
for( int j = 0; j < len_low_symbols; ++j )
for( i = 0; i < pos_states; ++i )
for( j = 0; j < len_low_symbols; ++j )
Bm_init( &len_encoder->bm_low[i][j] );
for( int i = 0; i < pos_states; ++i )
for( int j = 0; j < len_mid_symbols; ++j )
for( i = 0; i < pos_states; ++i )
for( j = 0; j < len_mid_symbols; ++j )
Bm_init( &len_encoder->bm_mid[i][j] );
for( int i = 0; i < len_high_symbols; ++i )
for( i = 0; i < len_high_symbols; ++i )
Bm_init( &len_encoder->bm_high[i] );
len_encoder->len_symbols = len_limit + 1 - min_match_len;
for( int i = 0; i < pos_states; ++i ) Lee_update_prices( len_encoder, i );
for( i = 0; i < pos_states; ++i ) Lee_update_prices( len_encoder, i );
}
void Lee_encode( struct Len_encoder * const len_encoder,
@ -433,16 +430,17 @@ struct Literal_encoder
Bit_model bm_literal[1<<literal_context_bits][0x300];
};
static inline void Lie_init( struct Literal_encoder * const literal_encoder )
{
for( int i = 0; i < 1<<literal_context_bits; ++i )
for( int j = 0; j < 0x300; ++j )
Bm_init( &literal_encoder->bm_literal[i][j] );
}
static inline int Lie_state( const int prev_byte )
{ return ( prev_byte >> ( 8 - literal_context_bits ) ); }
static inline void Lie_init( struct Literal_encoder * const literal_encoder )
{
int i, j;
for( i = 0; i < 1<<literal_context_bits; ++i )
for( j = 0; j < 0x300; ++j )
Bm_init( &literal_encoder->bm_literal[i][j] );
}
static inline void Lie_encode( struct Literal_encoder * const literal_encoder,
struct Range_encoder * const range_encoder,
uint8_t prev_byte, uint8_t symbol )
@ -450,29 +448,28 @@ static inline void Lie_encode( struct Literal_encoder * const literal_encoder,
static inline void Lie_encode_matched( struct Literal_encoder * const literal_encoder,
struct Range_encoder * const range_encoder,
uint8_t prev_byte, uint8_t match_byte, uint8_t symbol )
uint8_t prev_byte, uint8_t symbol, uint8_t match_byte )
{ Re_encode_matched( range_encoder, literal_encoder->bm_literal[Lie_state(prev_byte)], symbol, match_byte ); }
static inline int Lie_price_matched( struct Literal_encoder * const literal_encoder,
uint8_t prev_byte, uint8_t symbol, uint8_t match_byte )
{ return price_matched( literal_encoder->bm_literal[Lie_state(prev_byte)], symbol, match_byte ); }
static inline int Lie_price_symbol( struct Literal_encoder * const literal_encoder,
uint8_t prev_byte, uint8_t symbol )
{ return price_symbol( literal_encoder->bm_literal[Lie_state(prev_byte)], symbol, 8 ); }
static inline int Lie_price_matched( struct Literal_encoder * const literal_encoder,
uint8_t prev_byte, uint8_t symbol, uint8_t match_byte )
{ return price_matched( literal_encoder->bm_literal[Lie_state(prev_byte)], symbol, match_byte ); }
enum { lze_dis_align_mask = dis_align_size - 1,
lze_infinite_price = 0x0FFFFFFF,
lze_max_marker_size = 16,
num_rep_distances = 4 }; // must be 4
enum { infinite_price = 0x0FFFFFFF,
max_marker_size = 16,
num_rep_distances = 4 }; /* must be 4 */
struct Trial
{
State state;
int dis;
int prev_index; // index of prev trial in trials[]
int price; // dual use var; cumulative price, match length
int prev_index; /* index of prev trial in trials[] */
int price; /* dual use var; cumulative price, match length */
int reps[num_rep_distances];
};
@ -489,14 +486,14 @@ struct LZ_encoder
int longest_match_found;
uint32_t crc_;
Bit_model bm_match[St_states][pos_states];
Bit_model bm_rep[St_states];
Bit_model bm_rep0[St_states];
Bit_model bm_rep1[St_states];
Bit_model bm_rep2[St_states];
Bit_model bm_len[St_states][pos_states];
Bit_model bm_match[states][pos_states];
Bit_model bm_rep[states];
Bit_model bm_rep0[states];
Bit_model bm_rep1[states];
Bit_model bm_rep2[states];
Bit_model bm_len[states][pos_states];
Bit_model bm_dis_slot[max_dis_states][1<<dis_slot_bits];
Bit_model bm_dis[modeled_distances-end_dis_model];
Bit_model bm_dis[modeled_distances-end_dis_model+1];
Bit_model bm_align[dis_align_size];
struct Matchfinder * matchfinder;
@ -521,18 +518,19 @@ void LZe_fill_distance_prices( struct LZ_encoder * const encoder );
static inline uint32_t LZe_crc( struct LZ_encoder * const encoder )
{ return encoder->crc_ ^ 0xFFFFFFFFU; }
// move-to-front dis in/into reps
/* move-to-front dis in/into reps */
static inline void LZe_mtf_reps( const int dis, int reps[num_rep_distances] )
{
int i;
if( dis >= num_rep_distances )
{
for( int i = num_rep_distances - 1; i > 0; --i ) reps[i] = reps[i-1];
for( i = num_rep_distances - 1; i > 0; --i ) reps[i] = reps[i-1];
reps[0] = dis - num_rep_distances;
}
else if( dis > 0 )
{
const int distance = reps[dis];
for( int i = dis; i > 0; --i ) reps[i] = reps[i-1];
for( i = dis; i > 0; --i ) reps[i] = reps[i-1];
reps[0] = distance;
}
}
@ -546,9 +544,10 @@ static inline int LZe_price_rep_len1( struct LZ_encoder * const encoder,
static inline int LZe_price_rep( struct LZ_encoder * const encoder, const int rep,
const State state, const int pos_state )
{
int price;
if( rep == 0 ) return price0( encoder->bm_rep0[state] ) +
price1( encoder->bm_len[state][pos_state] );
int price = price1( encoder->bm_rep0[state] );
price = price1( encoder->bm_rep0[state] );
if( rep == 1 )
price += price0( encoder->bm_rep1[state] );
else
@ -559,27 +558,33 @@ static inline int LZe_price_rep( struct LZ_encoder * const encoder, const int re
return price;
}
static inline int LZe_price_pair( struct LZ_encoder * const encoder, const int dis,
const int len, const int pos_state )
static inline int LZe_price_pair( struct LZ_encoder * const encoder,
const int dis, const int len,
const int pos_state )
{
if( len <= min_match_len && dis >= modeled_distances )
return lze_infinite_price;
int price = Lee_price( &encoder->len_encoder, len, pos_state );
const int dis_state = get_dis_state( len );
int price;
if( len <= min_match_len && dis >= modeled_distances )
return infinite_price;
price = Lee_price( &encoder->len_encoder, len, pos_state );
if( dis < modeled_distances )
price += encoder->dis_prices[dis_state][dis];
else
price += encoder->dis_slot_prices[dis_state][get_slot( dis )] +
encoder->align_prices[dis & lze_dis_align_mask];
encoder->align_prices[dis & (dis_align_size - 1)];
return price;
}
static inline void LZe_encode_pair( struct LZ_encoder * const encoder,
const uint32_t dis, const int len, const int pos_state )
const uint32_t dis, const int len,
const int pos_state )
{
Lee_encode( &encoder->len_encoder, &encoder->range_encoder, len, pos_state );
const int dis_slot = get_slot( dis );
Re_encode_tree( &encoder->range_encoder, encoder->bm_dis_slot[get_dis_state(len)], dis_slot, dis_slot_bits );
Lee_encode( &encoder->len_encoder, &encoder->range_encoder, len, pos_state );
Re_encode_tree( &encoder->range_encoder,
encoder->bm_dis_slot[get_dis_state(len)],
dis_slot, dis_slot_bits );
if( dis_slot >= start_dis_model )
{
@ -607,16 +612,15 @@ static inline int LZe_read_match_distances( struct LZ_encoder * const encoder )
return len;
}
static inline bool LZe_move_pos( struct LZ_encoder * const encoder,
static inline void LZe_move_pos( struct LZ_encoder * const encoder,
int n, bool skip )
{
while( --n >= 0 )
{
if( skip ) skip = false;
else Mf_longest_match_len( encoder->matchfinder, 0 );
if( !Mf_move_pos( encoder->matchfinder ) ) return false;
Mf_move_pos( encoder->matchfinder );
}
return true;
}
static inline void LZe_backward( struct LZ_encoder * const encoder, int cur )
@ -626,7 +630,7 @@ static inline void LZe_backward( struct LZ_encoder * const encoder, int cur )
{
const int prev_index = encoder->trials[cur].prev_index;
struct Trial * const prev_trial = &encoder->trials[prev_index];
prev_trial->price = cur - prev_index; // len
prev_trial->price = cur - prev_index; /* len */
cur = *dis; *dis = prev_trial->dis; prev_trial->dis = cur;
cur = prev_index;
}

714
main.c
View file

@ -1,4 +1,4 @@
/* Clzip - A data compressor based on the LZMA algorithm
/* Clzip - Data compressor based on the LZMA algorithm
Copyright (C) 2010 Antonio Diaz Diaz.
This program is free software: you can redistribute it and/or modify
@ -35,6 +35,20 @@
#include <unistd.h>
#include <utime.h>
#include <sys/stat.h>
#if defined(__MSVCRT__)
#include <io.h>
#define fchmod(x,y) 0
#define fchown(x,y,z) 0
#define SIGHUP SIGTERM
#define S_ISSOCK(x) 0
#define S_IRGRP 0
#define S_IWGRP 0
#define S_IROTH 0
#define S_IWOTH 0
#endif
#if defined(__OS2__)
#include <io.h>
#endif
#include "carg_parser.h"
#include "clzip.h"
@ -55,11 +69,13 @@
#define ULLONG_MAX 0xFFFFFFFFFFFFFFFFULL
#endif
long long int llabs( long long int number );
const char * const Program_name = "Clzip";
const char * const program_name = "clzip";
const char * const program_year = "2010";
const char * invocation_name = 0;
const char * const Program_name = "Clzip";
const char * const program_name = "clzip";
const char * const program_year = "2010";
#ifdef O_BINARY
const int o_binary = O_BINARY;
@ -74,19 +90,20 @@ struct { const char * from; const char * to; } const known_extensions[] = {
struct Lzma_options
{
int dictionary_size; // 4KiB..512MiB
int match_len_limit; // 5..273
int dictionary_size; /* 4KiB..512MiB */
int match_len_limit; /* 5..273 */
};
enum Mode { m_compress = 0, m_decompress, m_test };
enum Mode { m_compress, m_decompress, m_test };
char * output_filename = 0;
int outfd = -1;
int verbosity = 0;
mode_t outfd_mode = S_IRUSR | S_IWUSR;
bool delete_output_on_interrupt = false;
// assure at least a minimum size for buffer `buf'
/* assure at least a minimum size for buffer `buf' */
inline void * resize_buffer( void * buf, const int min_size )
{
if( buf ) buf = realloc( buf, min_size );
@ -97,7 +114,7 @@ inline void * resize_buffer( void * buf, const int min_size )
static void show_help()
{
printf( "%s - A data compressor based on the LZMA algorithm.\n", Program_name );
printf( "%s - Data compressor based on the LZMA algorithm.\n", Program_name );
printf( "\nUsage: %s [options] [files]\n", invocation_name );
printf( "\nOptions:\n" );
printf( " -h, --help display this help and exit\n" );
@ -136,39 +153,33 @@ static void show_version()
}
static const char * format_num( long long num, long long limit,
const int set_prefix )
static const char * format_num( long long num )
{
const char * const si_prefix[8] =
{ "k", "M", "G", "T", "P", "E", "Z", "Y" };
const char * const binary_prefix[8] =
const char * const prefix[8] =
{ "Ki", "Mi", "Gi", "Ti", "Pi", "Ei", "Zi", "Yi" };
static bool si = false;
static char buf[16];
if( set_prefix ) si = ( set_prefix > 0 );
const int factor = ( si ) ? 1000 : 1024;
const char * const *prefix = ( si ) ? si_prefix : binary_prefix;
enum { buf_size = 16, factor = 1024 };
static char buf[buf_size];
const char *p = "";
limit = max( 999LL, min( 999999LL, limit ) );
int i;
for( int i = 0; i < 8 && ( llabs( num ) > limit ||
for( i = 0; i < 8 && ( llabs( num ) > 9999 ||
( llabs( num ) >= factor && num % factor == 0 ) ); ++i )
{ num /= factor; p = prefix[i]; }
snprintf( buf, sizeof buf, "%lld %s", num, p );
snprintf( buf, buf_size, "%lld %s", num, p );
return buf;
}
static long long getnum( const char * const ptr, const int bs,
static long long getnum( const char * const ptr,
const long long llimit, const long long ulimit )
{
errno = 0;
long long result;
char *tail;
long long result = strtoll( ptr, &tail, 0 );
errno = 0;
result = strtoll( ptr, &tail, 0 );
if( tail == ptr )
{
show_error( "bad or missing numerical argument", 0, true );
show_error( "Bad or missing numerical argument.", 0, true );
exit( 1 );
}
@ -176,13 +187,11 @@ static long long getnum( const char * const ptr, const int bs,
{
int factor = ( tail[1] == 'i' ) ? 1024 : 1000;
int exponent = 0;
int i;
bool bad_multiplier = false;
switch( tail[0] )
{
case ' ': break;
case 'b': if( bs > 0 ) { factor = bs; exponent = 1; }
else bad_multiplier = true;
break;
case 'Y': exponent = 8; break;
case 'Z': exponent = 7; break;
case 'E': exponent = 6; break;
@ -198,10 +207,10 @@ static long long getnum( const char * const ptr, const int bs,
}
if( bad_multiplier )
{
show_error( "bad multiplier in numerical argument", 0, true );
show_error( "Bad multiplier in numerical argument.", 0, true );
exit( 1 );
}
for( int i = 0; i < exponent; ++i )
for( i = 0; i < exponent; ++i )
{
if( LLONG_MAX / factor >= llabs( result ) ) result *= factor;
else { errno = ERANGE; break; }
@ -210,7 +219,7 @@ static long long getnum( const char * const ptr, const int bs,
if( !errno && ( result < llimit || result > ulimit ) ) errno = ERANGE;
if( errno )
{
show_error( "numerical argument out of limits", 0, false );
show_error( "Numerical argument out of limits.", 0, false );
exit( 1 );
}
return result;
@ -224,13 +233,14 @@ static int get_dict_size( const char * const arg )
if( bits >= min_dictionary_bits &&
bits <= max_dictionary_bits && *tail == 0 )
return ( 1 << bits );
return getnum( arg, 0, min_dictionary_size, max_dictionary_size );
return getnum( arg, min_dictionary_size, max_dictionary_size );
}
static int extension_index( const char * const name )
{
for( int i = 0; known_extensions[i].from; ++i )
int i;
for( i = 0; known_extensions[i].from; ++i )
{
const char * const ext = known_extensions[i].from;
if( strlen( name ) > strlen( ext ) &&
@ -249,7 +259,7 @@ static int open_instream( const char * const name, struct stat * const in_statsp
if( program_mode == m_compress && !force && eindex >= 0 )
{
if( verbosity >= 0 )
fprintf( stderr, "%s: input file `%s' already has `%s' suffix.\n",
fprintf( stderr, "%s: Input file `%s' already has `%s' suffix.\n",
program_name, name, known_extensions[eindex].from );
}
else
@ -265,14 +275,16 @@ static int open_instream( const char * const name, struct stat * const in_statsp
{
const int i = fstat( infd, in_statsp );
const mode_t mode = in_statsp->st_mode;
if( i < 0 || !( S_ISREG( mode ) || ( to_stdout &&
( S_ISFIFO( mode ) || S_ISSOCK( mode ) ||
S_ISBLK( mode ) || S_ISCHR( mode ) ) ) ) )
const bool can_read = ( i == 0 &&
( S_ISBLK( mode ) || S_ISCHR( mode ) ||
S_ISFIFO( mode ) || S_ISSOCK( mode ) ) );
if( i != 0 || ( !S_ISREG( mode ) && ( !to_stdout || !can_read ) ) )
{
if( verbosity >= 0 )
fprintf( stderr, "%s: input file `%s' is not a regular file%s.\n",
fprintf( stderr, "%s: Input file `%s' is not a regular file%s.\n",
program_name, name,
to_stdout ? "" : " and `--stdout' was not specified" );
( can_read && !to_stdout ) ?
" and `--stdout' was not specified" : "" );
close( infd );
infd = -1;
}
@ -311,7 +323,7 @@ static void set_d_outname( const char * const name, const int i )
strcpy( output_filename, name );
strcat( output_filename, ".out" );
if( verbosity >= 0 )
fprintf( stderr, "%s: can't guess original name for `%s' -- using `%s'.\n",
fprintf( stderr, "%s: Can't guess original name for `%s' -- using `%s'.\n",
program_name, name, output_filename );
}
@ -322,18 +334,14 @@ static bool open_outstream( const bool force )
if( force ) flags |= O_TRUNC; else flags |= O_EXCL;
outfd = open( output_filename, flags, outfd_mode );
if( outfd < 0 )
if( outfd < 0 && verbosity >= 0 )
{
if( errno == EEXIST ) outfd = -2; else outfd = -1;
if( verbosity >= 0 )
{
if( outfd == -2 )
fprintf( stderr, "%s: Output file %s already exists, skipping.\n",
program_name, output_filename );
else
fprintf( stderr, "%s: Can't create output file `%s': %s.\n",
program_name, output_filename, strerror( errno ) );
}
if( errno == EEXIST )
fprintf( stderr, "%s: Output file %s already exists, skipping.\n",
program_name, output_filename );
else
fprintf( stderr, "%s: Can't create output file `%s': %s.\n",
program_name, output_filename, strerror( errno ) );
}
return ( outfd >= 0 );
}
@ -341,7 +349,7 @@ static bool open_outstream( const bool force )
static bool check_tty( const int infd, const enum Mode program_mode )
{
if( program_mode == m_compress && isatty( outfd ) )
if( program_mode == m_compress && outfd >= 0 && isatty( outfd ) )
{
show_error( "I won't write compressed data to a terminal.", 0, true );
return false;
@ -356,246 +364,6 @@ static bool check_tty( const int infd, const enum Mode program_mode )
}
// Set permissions, owner and times.
static void close_and_set_permissions( const struct stat * const in_statsp )
{
bool error = false;
if( in_statsp )
{
if( fchmod( outfd, in_statsp->st_mode ) != 0 ) error = true;
else (void)fchown( outfd, in_statsp->st_uid, in_statsp->st_gid );
// fchown will in many cases return with EPERM, which can be safely ignored.
}
if( close( outfd ) == 0 ) outfd = -1;
else cleanup_and_fail( 1 );
delete_output_on_interrupt = false;
if( !in_statsp ) return;
if( !error )
{
struct utimbuf t;
t.actime = in_statsp->st_atime;
t.modtime = in_statsp->st_mtime;
if( utime( output_filename, &t ) != 0 ) error = true;
}
if( error )
{
show_error( "I can't change output file attributes.", 0, false );
cleanup_and_fail( 1 );
}
}
static bool next_filename()
{
const unsigned int len = strlen( known_extensions[0].from );
if( strlen( output_filename ) >= len + 5 ) // "*00001.lz"
for( int i = strlen( output_filename ) - len - 1, j = 0; j < 5; --i, ++j )
{
if( output_filename[i] < '9' ) { ++output_filename[i]; return true; }
else output_filename[i] = '0';
}
return false;
}
static int compress( const long long member_size, const long long volume_size,
const struct Lzma_options * const encoder_options,
const int infd, struct Pretty_print * const pp,
const struct stat * const in_statsp )
{
if( verbosity >= 1 ) Pp_show_msg( pp, 0 );
File_header header;
Fh_set_magic( header );
if( !Fh_set_dictionary_size( header, encoder_options->dictionary_size ) ||
encoder_options->match_len_limit < min_match_len_limit ||
encoder_options->match_len_limit > max_match_len )
internal_error( "invalid argument to encoder" );
struct Matchfinder matchfinder;
Mf_init( &matchfinder, Fh_get_dictionary_size( header ),
encoder_options->match_len_limit, infd );
Fh_set_dictionary_size( header, Mf_dictionary_size( &matchfinder ) );
long long in_size = 0, out_size = 0, partial_volume_size = 0;
int retval = 0;
while( true ) // encode one member per iteration
{
struct LZ_encoder encoder;
LZe_init( &encoder, &matchfinder, header, outfd );
const long long size =
min( member_size, volume_size - partial_volume_size );
if( !LZe_encode_member( &encoder, size ) )
{ Pp_show_msg( pp, "encoder error" ); retval = 1; break; }
in_size += Mf_data_position( &matchfinder );
out_size += LZe_member_position( &encoder );
partial_volume_size += LZe_member_position( &encoder );
LZe_free( &encoder );
if( Mf_finished( &matchfinder ) ) break;
if( partial_volume_size >= volume_size - min_dictionary_size )
{
partial_volume_size = 0;
if( delete_output_on_interrupt )
{
close_and_set_permissions( in_statsp );
if( !next_filename() )
{ Pp_show_msg( pp, "too many volume files" ); retval = 1; break; }
if( !open_outstream( true ) ) { retval = 1; break; }
delete_output_on_interrupt = true;
}
}
if( !Mf_reset( &matchfinder ) )
{ Pp_show_msg( pp, 0 );
show_error( "can't reset matchfinder", errno, false );
retval = 1; break; }
}
if( retval == 0 && verbosity >= 1 )
{
if( in_size <= 0 || out_size <= 0 )
fprintf( stderr, "no data compressed.\n" );
else
fprintf( stderr, "%6.3f:1, %6.3f bits/byte, "
"%5.2f%% saved, %lld in, %lld out.\n",
(double)in_size / out_size,
( 8.0 * out_size ) / in_size,
100.0 * ( 1.0 - ( (double)out_size / in_size ) ),
in_size, out_size );
}
Mf_free( &matchfinder );
return retval;
}
static int decompress( const int infd, struct Pretty_print * const pp,
const bool testing )
{
struct Range_decoder rdec;
Rd_init( &rdec, infd );
long long partial_file_pos = 0;
int retval = 0;
for( bool first_member = true; ; first_member = false, Pp_reset( pp ) )
{
File_header header;
Rd_reset_member_position( &rdec );
for( int i = 0; i < Fh_size; ++i )
header[i] = Rd_get_byte( &rdec );
if( Rd_finished( &rdec ) ) // End Of File
{
if( first_member )
{ Pp_show_msg( pp, "error reading member header" ); retval = 1; }
break;
}
if( !Fh_verify_magic( header ) )
{
if( !first_member ) break; // trailing garbage
Pp_show_msg( pp, "bad magic number (file not in lzip format)" );
retval = 2; break;
}
if( !Fh_verify_version( header ) )
{
if( verbosity >= 0 )
{ Pp_show_msg( pp, 0 );
fprintf( stderr, "version %d member format not supported, newer %s needed.\n",
Fh_version( header ), program_name ); }
retval = 2; break;
}
if( Fh_get_dictionary_size( header ) < min_dictionary_size ||
Fh_get_dictionary_size( header ) > max_dictionary_size )
{ Pp_show_msg( pp, "invalid dictionary size in member header" );
retval = 2; break; }
if( verbosity >= 1 )
{
Pp_show_msg( pp, 0 );
if( verbosity >= 2 )
fprintf( stderr, "version %d, dictionary size %7sB. ",
Fh_version( header ),
format_num( Fh_get_dictionary_size( header ), 9999, 0 ) );
}
struct LZ_decoder decoder;
LZd_init( &decoder, header, &rdec, outfd );
const int result = LZd_decode_member( &decoder, pp );
partial_file_pos += Rd_member_position( &rdec );
LZd_free( &decoder );
if( result != 0 )
{
if( verbosity >= 0 && result <= 2 )
{
Pp_show_msg( pp, 0 );
if( result == 2 )
fprintf( stderr, "file ends unexpectedly at pos %lld\n",
partial_file_pos );
else
fprintf( stderr, "decoder error at pos %lld\n", partial_file_pos );
}
retval = 2; break;
}
if( verbosity >= 1 )
{ if( testing ) fprintf( stderr, "ok\n" );
else fprintf( stderr, "done\n" ); }
}
Rd_free( &rdec );
return retval;
}
void signal_handler( int sig )
{
sig = 0; // keep compiler happy
show_error( "Control-C or similar caught, quitting.", 0, false );
cleanup_and_fail( 1 );
}
static void set_signals()
{
signal( SIGHUP, signal_handler );
signal( SIGINT, signal_handler );
signal( SIGTERM, signal_handler );
}
int verbosity = 0;
void Pp_init( struct Pretty_print * const pp, const char * const filenames[],
const int num_filenames )
{
pp->name_ = 0;
pp->stdin_name = "(stdin)";
pp->longest_name = 0;
pp->first_post = false;
unsigned int stdin_name_len = strlen( pp->stdin_name );
for( int i = 0; i < num_filenames; ++i )
{
const char * const s = filenames[i];
const int len = ( !strcmp( s, "-" ) ? stdin_name_len : strlen( s ) );
if( len > pp->longest_name ) pp->longest_name = len;
}
if( pp->longest_name == 0 ) pp->longest_name = stdin_name_len;
}
void Pp_show_msg( struct Pretty_print * const pp, const char * const msg )
{
if( verbosity >= 0 )
{
if( pp->first_post )
{
pp->first_post = false;
fprintf( stderr, " %s: ", pp->name_ );
const int len = pp->longest_name - strlen( pp->name_ );
for( int i = 0; i < len; ++i ) fprintf( stderr, " " );
if( !msg ) fflush( stderr );
}
if( msg ) fprintf( stderr, "%s.\n", msg );
}
}
void cleanup_and_fail( const int retval )
{
if( delete_output_on_interrupt )
@ -612,17 +380,260 @@ void cleanup_and_fail( const int retval )
}
/* Set permissions, owner and times. */
static void close_and_set_permissions( const struct stat * const in_statsp )
{
bool error = false;
if( in_statsp )
{
if( fchmod( outfd, in_statsp->st_mode ) != 0 ||
( fchown( outfd, in_statsp->st_uid, in_statsp->st_gid ) != 0 &&
errno != EPERM ) ) error = true;
/* fchown will in many cases return with EPERM, which can be safely ignored. */
}
if( close( outfd ) == 0 ) outfd = -1;
else cleanup_and_fail( 1 );
delete_output_on_interrupt = false;
if( !in_statsp ) return;
if( !error )
{
struct utimbuf t;
t.actime = in_statsp->st_atime;
t.modtime = in_statsp->st_mtime;
if( utime( output_filename, &t ) != 0 ) error = true;
}
if( error )
{
show_error( "Can't change output file attributes.", 0, false );
cleanup_and_fail( 1 );
}
}
static bool next_filename()
{
const unsigned int len = strlen( known_extensions[0].from );
int i, j;
if( strlen( output_filename ) >= len + 5 ) /* "*00001.lz" */
for( i = strlen( output_filename ) - len - 1, j = 0; j < 5; --i, ++j )
{
if( output_filename[i] < '9' ) { ++output_filename[i]; return true; }
else output_filename[i] = '0';
}
return false;
}
static int compress( const long long member_size, const long long volume_size,
const struct Lzma_options * const encoder_options,
const int infd, struct Pretty_print * const pp,
const struct stat * const in_statsp )
{
long long in_size = 0, out_size = 0, partial_volume_size = 0;
int retval = 0;
File_header header;
struct Matchfinder matchfinder;
if( verbosity >= 1 ) Pp_show_msg( pp, 0 );
Fh_set_magic( header );
if( !Fh_set_dictionary_size( header, encoder_options->dictionary_size ) ||
encoder_options->match_len_limit < min_match_len_limit ||
encoder_options->match_len_limit > max_match_len )
internal_error( "invalid argument to encoder" );
Mf_init( &matchfinder, Fh_get_dictionary_size( header ),
encoder_options->match_len_limit, infd );
Fh_set_dictionary_size( header, Mf_dictionary_size( &matchfinder ) );
while( true ) /* encode one member per iteration */
{
struct LZ_encoder encoder;
const long long size =
min( member_size, volume_size - partial_volume_size );
LZe_init( &encoder, &matchfinder, header, outfd );
if( !LZe_encode_member( &encoder, size ) )
{ Pp_show_msg( pp, "Encoder error" ); retval = 1; break; }
in_size += Mf_data_position( &matchfinder );
out_size += LZe_member_position( &encoder );
partial_volume_size += LZe_member_position( &encoder );
LZe_free( &encoder );
if( Mf_finished( &matchfinder ) ) break;
if( partial_volume_size >= volume_size - min_dictionary_size )
{
partial_volume_size = 0;
if( delete_output_on_interrupt )
{
close_and_set_permissions( in_statsp );
if( !next_filename() )
{ Pp_show_msg( pp, "Too many volume files" ); retval = 1; break; }
if( !open_outstream( true ) ) { retval = 1; break; }
delete_output_on_interrupt = true;
}
}
Mf_reset( &matchfinder );
}
if( retval == 0 && verbosity >= 1 )
{
if( in_size <= 0 || out_size <= 0 )
fprintf( stderr, "No data compressed.\n" );
else
fprintf( stderr, "%6.3f:1, %6.3f bits/byte, "
"%5.2f%% saved, %lld in, %lld out.\n",
(double)in_size / out_size,
( 8.0 * out_size ) / in_size,
100.0 * ( 1.0 - ( (double)out_size / in_size ) ),
in_size, out_size );
}
Mf_free( &matchfinder );
return retval;
}
static int decompress( const int infd, struct Pretty_print * const pp,
const bool testing )
{
long long partial_file_pos = 0;
struct Range_decoder rdec;
int retval = 0, i, result;
bool first_member;
Rd_init( &rdec, infd );
for( first_member = true; ; first_member = false, Pp_reset( pp ) )
{
File_header header;
struct LZ_decoder decoder;
Rd_reset_member_position( &rdec );
for( i = 0; i < Fh_size; ++i )
header[i] = Rd_get_byte( &rdec );
if( Rd_finished( &rdec ) ) /* End Of File */
{
if( first_member )
{ Pp_show_msg( pp, "Error reading member header" ); retval = 1; }
break;
}
if( !Fh_verify_magic( header ) )
{
if( !first_member ) break; /* trailing garbage */
Pp_show_msg( pp, "Bad magic number (file not in lzip format)" );
retval = 2; break;
}
if( !Fh_verify_version( header ) )
{
if( verbosity >= 0 )
{ Pp_show_msg( pp, 0 );
fprintf( stderr, "Version %d member format not supported.\n",
Fh_version( header ) ); }
retval = 2; break;
}
if( Fh_get_dictionary_size( header ) < min_dictionary_size ||
Fh_get_dictionary_size( header ) > max_dictionary_size )
{ Pp_show_msg( pp, "Invalid dictionary size in member header" );
retval = 2; break; }
if( verbosity >= 1 )
{
Pp_show_msg( pp, 0 );
if( verbosity >= 2 )
fprintf( stderr, "version %d, dictionary size %7sB. ",
Fh_version( header ),
format_num( Fh_get_dictionary_size( header ) ) );
}
LZd_init( &decoder, header, &rdec, outfd );
result = LZd_decode_member( &decoder, pp );
partial_file_pos += Rd_member_position( &rdec );
LZd_free( &decoder );
if( result != 0 )
{
if( verbosity >= 0 && result <= 2 )
{
Pp_show_msg( pp, 0 );
if( result == 2 )
fprintf( stderr, "File ends unexpectedly at pos %lld\n",
partial_file_pos );
else
fprintf( stderr, "Decoder error at pos %lld\n", partial_file_pos );
}
retval = 2; break;
}
if( verbosity >= 1 )
{ if( testing ) fprintf( stderr, "ok\n" );
else fprintf( stderr, "done\n" ); }
}
Rd_free( &rdec );
return retval;
}
void signal_handler( int sig )
{
sig = 0; /* keep compiler happy */
show_error( "Control-C or similar caught, quitting.", 0, false );
cleanup_and_fail( 1 );
}
static void set_signals()
{
signal( SIGHUP, signal_handler );
signal( SIGINT, signal_handler );
signal( SIGTERM, signal_handler );
}
void Pp_init( struct Pretty_print * const pp, const char * const filenames[],
const int num_filenames, const int v )
{
unsigned int stdin_name_len;
int i;
pp->name_ = 0;
pp->stdin_name = "(stdin)";
pp->longest_name = 0;
pp->verbosity = v;
pp->first_post = false;
stdin_name_len = strlen( pp->stdin_name );
for( i = 0; i < num_filenames; ++i )
{
const char * const s = filenames[i];
const int len = ( !strcmp( s, "-" ) ? stdin_name_len : strlen( s ) );
if( len > pp->longest_name ) pp->longest_name = len;
}
if( pp->longest_name == 0 ) pp->longest_name = stdin_name_len;
}
void Pp_show_msg( struct Pretty_print * const pp, const char * const msg )
{
if( verbosity >= 0 )
{
if( pp->first_post )
{
int i, len;
pp->first_post = false;
fprintf( stderr, " %s: ", pp->name_ );
len = pp->longest_name - strlen( pp->name_ );
for( i = 0; i < len; ++i ) fprintf( stderr, " " );
if( !msg ) fflush( stderr );
}
if( msg ) fprintf( stderr, "%s.\n", msg );
}
}
void show_error( const char * const msg, const int errcode, const bool help )
{
if( verbosity >= 0 )
{
if( msg && msg[0] != 0 )
if( msg && msg[0] )
{
fprintf( stderr, "%s: %s", program_name, msg );
if( errcode > 0 ) fprintf( stderr, ": %s", strerror( errcode ) );
fprintf( stderr, "\n" );
}
if( help && invocation_name && invocation_name[0] != 0 )
if( help && invocation_name && invocation_name[0] )
fprintf( stderr, "Try `%s --help' for more information.\n", invocation_name );
}
}
@ -630,27 +641,24 @@ void show_error( const char * const msg, const int errcode, const bool help )
void internal_error( const char * const msg )
{
const char * const e = "internal error: ";
char * s = resize_buffer( 0, strlen( e ) + strlen( msg ) + 1 );
strcpy( s, e );
strcat( s, msg );
show_error( s, 0, false );
free( s );
if( verbosity >= 0 )
fprintf( stderr, "%s: internal error: %s.\n", program_name, msg );
exit( 3 );
}
// Returns the number of bytes really read.
// If (returned value < size) and (errno == 0), means EOF was reached.
//
/* Returns the number of bytes really read.
If (returned value < size) and (errno == 0), means EOF was reached.
*/
int readblock( const int fd, uint8_t * const buf, const int size )
{
int rest = size;
errno = 0;
while( rest > 0 )
{
int n;
errno = 0;
const int n = read( fd, buf + size - rest, rest );
n = read( fd, buf + size - rest, rest );
if( n > 0 ) rest -= n;
else if( n == 0 ) break;
else if( errno != EINTR && errno != EAGAIN ) break;
@ -659,17 +667,18 @@ int readblock( const int fd, uint8_t * const buf, const int size )
}
// Returns the number of bytes really written.
// If (returned value < size), it is always an error.
//
/* Returns the number of bytes really written.
If (returned value < size), it is always an error.
*/
int writeblock( const int fd, const uint8_t * const buf, const int size )
{
int rest = size;
errno = 0;
while( rest > 0 )
{
int n;
errno = 0;
const int n = write( fd, buf + size - rest, rest );
n = write( fd, buf + size - rest, rest );
if( n > 0 ) rest -= n;
else if( errno && errno != EINTR && errno != EAGAIN ) break;
}
@ -677,29 +686,31 @@ int writeblock( const int fd, const uint8_t * const buf, const int size )
}
CRC32 crc32;
int main( const int argc, const char * const argv[] )
{
// Mapping from gzip/bzip2 style 1..9 compression modes
// to the corresponding LZMA compression modes.
/* Mapping from gzip/bzip2 style 1..9 compression modes
to the corresponding LZMA compression modes. */
const struct Lzma_options option_mapping[] =
{
{ 1 << 16, 5 }, // -0
{ 1 << 20, 10 }, // -1
{ 3 << 19, 12 }, // -2
{ 1 << 21, 17 }, // -3
{ 3 << 20, 26 }, // -4
{ 1 << 22, 44 }, // -5
{ 1 << 23, 80 }, // -6
{ 1 << 24, 108 }, // -7
{ 3 << 23, 163 }, // -8
{ 1 << 25, 273 } }; // -9
struct Lzma_options encoder_options = option_mapping[6]; // default = "-6"
{ 1 << 20, 5 }, /* -0 */
{ 1 << 20, 5 }, /* -1 */
{ 3 << 19, 6 }, /* -2 */
{ 1 << 21, 8 }, /* -3 */
{ 3 << 20, 12 }, /* -4 */
{ 1 << 22, 20 }, /* -5 */
{ 1 << 23, 36 }, /* -6 */
{ 1 << 24, 68 }, /* -7 */
{ 3 << 23, 132 }, /* -8 */
{ 1 << 25, 273 } }; /* -9 */
struct Lzma_options encoder_options = option_mapping[6]; /* default = "-6" */
long long member_size = LLONG_MAX;
long long volume_size = LLONG_MAX;
int infd = -1;
enum Mode program_mode = m_compress;
int argind = 0;
int retval = 0;
int i;
bool filenames_given = false;
bool force = false;
bool keep_input_files = false;
bool to_stdout = false;
@ -707,13 +718,12 @@ int main( const int argc, const char * const argv[] )
const char * default_output_filename = "";
const char ** filenames = 0;
int num_filenames = 0;
invocation_name = argv[0];
CRC32_init();
struct Pretty_print pp;
const struct ap_Option options[] =
{
{ '0', 0, ap_no },
{ '1', "fast", ap_no },
{ '0', "fast", ap_no },
{ '1', 0, ap_no },
{ '2', 0, ap_no },
{ '3', 0, ap_no },
{ '4', 0, ap_no },
@ -740,44 +750,51 @@ int main( const int argc, const char * const argv[] )
{ 0 , 0, ap_no } };
struct Arg_parser parser;
invocation_name = argv[0];
CRC32_init();
if( !ap_init( &parser, argc, argv, options, 0 ) )
{ show_error( "memory exhausted", 0, false ); return 1; }
if( ap_error( &parser ) ) // bad option
{ show_error( "Memory exhausted.", 0, false ); return 1; }
if( ap_error( &parser ) ) /* bad option */
{ show_error( ap_error( &parser ), 0, true ); return 1; }
int argind = 0;
for( ; argind < ap_arguments( &parser ); ++argind )
{
const int code = ap_code( &parser, argind );
const char * const arg = ap_argument( &parser, argind );
if( !code ) break; // no more options
if( !code ) break; /* no more options */
switch( code )
{
case '0': case '1': case '2': case '3': case '4':
case '0':
case '1': case '2': case '3': case '4':
case '5': case '6': case '7': case '8': case '9':
encoder_options = option_mapping[code-'0']; break;
case 'b': member_size = getnum( arg, 0, 100000, LLONG_MAX / 2 ); break;
case 'b': member_size = getnum( arg, 100000, LLONG_MAX / 2 ); break;
case 'c': to_stdout = true; break;
case 'd': program_mode = m_decompress; break;
case 'e': break; // ignored by now
case 'e': break; /* ignored by now */
case 'f': force = true; break;
case 'h': show_help(); return 0;
case 'k': keep_input_files = true; break;
case 'm': encoder_options.match_len_limit =
getnum( arg, 0, min_match_len_limit, max_match_len ); break;
getnum( arg, min_match_len_limit, max_match_len ); break;
case 'o': default_output_filename = arg; break;
case 'q': verbosity = -1; break;
case 's': encoder_options.dictionary_size = get_dict_size( arg );
break;
case 'S': volume_size = getnum( arg, 0, 100000, LLONG_MAX / 2 ); break;
case 'S': volume_size = getnum( arg, 100000, LLONG_MAX / 2 ); break;
case 't': program_mode = m_test; break;
case 'v': if( verbosity < 4 ) ++verbosity; break;
case 'V': show_version(); return 0;
default : internal_error( "uncaught option" );
}
}
} /* end process options */
#if defined(__MSVCRT__) || defined(__OS2__)
_setmode( STDIN_FILENO, O_BINARY );
_setmode( STDOUT_FILENO, O_BINARY );
#endif
bool filenames_given = false;
for( ; argind < ap_arguments( &parser ); ++argind )
{
if( strcmp( ap_argument( &parser, argind ), "-" ) )
@ -797,8 +814,7 @@ int main( const int argc, const char * const argv[] )
( filenames_given || default_output_filename[0] ) )
set_signals();
struct Pretty_print pp;
Pp_init( &pp, filenames, num_filenames );
Pp_init( &pp, filenames, num_filenames, verbosity );
if( program_mode == m_test )
outfd = -1;
else if( program_mode == m_compress )
@ -808,10 +824,11 @@ int main( const int argc, const char * const argv[] )
}
output_filename = resize_buffer( output_filename, 1 );
int retval = 0;
for( int i = 0; i < num_filenames; ++i )
for( i = 0; i < num_filenames; ++i )
{
int tmp;
struct stat in_stats;
const struct stat * in_statsp;
output_filename[0] = 0;
if( !filenames[i][0] || !strcmp( filenames[i], "-" ) )
@ -844,8 +861,8 @@ int main( const int argc, const char * const argv[] )
}
else
{
const int eindex = extension_index( filenames[i] );
input_filename = filenames[i];
const int eindex = extension_index( input_filename );
infd = open_instream( input_filename, &in_stats, program_mode,
eindex, force, to_stdout );
if( infd < 0 ) { if( retval < 1 ) retval = 1; continue; }
@ -872,9 +889,8 @@ int main( const int argc, const char * const argv[] )
if( output_filename[0] && !to_stdout && program_mode != m_test )
delete_output_on_interrupt = true;
const struct stat * const in_statsp = input_filename[0] ? &in_stats : 0;
in_statsp = input_filename[0] ? &in_stats : 0;
Pp_set_name( &pp, input_filename );
int tmp = 0;
if( program_mode == m_compress )
tmp = compress( member_size, volume_size, &encoder_options, infd,
&pp, in_statsp );
@ -894,9 +910,7 @@ int main( const int argc, const char * const argv[] )
}
if( outfd >= 0 && close( outfd ) != 0 )
{
if( verbosity >= 0 )
fprintf( stderr, "%s: Can't close stdout: %s.\n",
program_name, strerror( errno ) );
show_error( "Can't close stdout", errno, false );
if( retval < 1 ) retval = 1;
}
free( output_filename );

24
testsuite/check.sh
View file

@ -1,5 +1,5 @@
#! /bin/sh
# check script for Clzip - A data compressor based on the LZMA algorithm
# check script for Clzip - Data compressor based on the LZMA algorithm
# Copyright (C) 2010 Antonio Diaz Diaz.
#
# This script is free software: you have unlimited permission
@ -19,14 +19,23 @@ fi
if [ -d tmp ] ; then rm -rf tmp ; fi
mkdir tmp
printf "testing clzip..."
printf "testing clzip-%s..." "$2"
cd "${objdir}"/tmp
cat "${testdir}"/test1 > in || framework_failure
cat "${testdir}"/test.txt > in || framework_failure
fail=0
"${LZIP}" -cd "${testdir}"/test1.lz > copy || fail=1
"${LZIP}" -t "${testdir}"/test_v0.lz || fail=1
printf .
"${LZIP}" -cd "${testdir}"/test_v0.lz > copy || fail=1
cmp in copy || fail=1
printf .
"${LZIP}" -t "${testdir}"/test_v1.lz || fail=1
printf .
"${LZIP}" -cd "${testdir}"/test_v1.lz > copy || fail=1
cmp in copy || fail=1
printf .
for i in s4Ki 0 1 2 3 4 5 6 7 8 9 ; do
"${LZIP}" -k -$i in || fail=1
@ -53,12 +62,17 @@ for i in s4Ki 0 1 2 3 4 5 6 7 8 9 ; do
done
for i in s4Ki 0 1 2 3 4 5 6 7 8 9 ; do
"${LZIP}" -fe -$i -o out < in || fail=1
"${LZIP}" -f -$i -o out < in || fail=1
"${LZIP}" -df -o copy < out.lz || fail=1
cmp in copy || fail=1
printf .
done
"${LZIP}" -$i < in > anyothername || fail=1
"${LZIP}" -dq anyothername || fail=1
cmp in anyothername.out || fail=1
printf .
echo
if [ ${fail} = 0 ] ; then
echo "tests completed successfully."

0
testsuite/test1 → testsuite/test.txt
View file

0
testsuite/test1.lz → testsuite/test_v0.lz
View file

BIN
testsuite/test_v1.lz Normal file
View file

Binary file not shown.